
%%html
<!-- Style properties for markdown cells throughout the project -->

<style>
.text
{    
    font-family:Verdana;
    font-size: 13px;
}
.text11
{
    font-family:Verdana;
    font-size:11px;
}
.text12
{    
    font-family:Verdana;
    font-size: 12px;
}
.text14
{    
    font-family:Verdana;
    font-size: 14px;
}
.text16
{
    font-family:Verdana;
    font-size: 16px;
}
.jump
{
    font-size: 12px;
}
.expand
{
    display:list-item
}
</style>


#30 August 2021 16:30 IST

STATIC = True


import datetime
import numpy as np
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
import plotly.io as pio
import warnings
from plotly.subplots import make_subplots
from collections import Counter
from geopy.geocoders import Nominatim
from plotly.colors import n_colors
from scipy.stats import chisquare
from tabulate import tabulate

# pio.renderers.default='notebook'
warnings.filterwarnings('ignore')

mapbox_personal_token = 'pk.eyJ1IjoidnlhZHV2YW5zaGkiLCJhIjoiY2txYWxteGprMG52NDJvam8wcTZmYTc4cyJ9.cvg0lKU2zkXKCLq1mWjflg'


#Reading the data

if STATIC:
    df = pd.read_csv('static_raw_df.csv')
else:
    df = pd.read_csv('https://raw.githubusercontent.com/owid/covid-19-data/master/public/data/owid-covid-data.csv')

df.head()


# #Saving to local

# df.to_csv('static_raw_df.csv', index=False)


df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 112864 entries, 0 to 112863
Data columns (total 62 columns):
 #   Column                                 Non-Null Count   Dtype  
---  ------                                 --------------   -----  
 0   iso_code                               112864 non-null  object 
 1   continent                              107664 non-null  object 
 2   location                               112864 non-null  object 
 3   date                                   112864 non-null  object 
 4   total_cases                            107847 non-null  float64
 5   new_cases                              107844 non-null  float64
 6   new_cases_smoothed                     106829 non-null  float64
 7   total_deaths                           97300 non-null   float64
 8   new_deaths                             97455 non-null   float64
 9   new_deaths_smoothed                    106829 non-null  float64
 10  total_cases_per_million                107277 non-null  float64
 11  new_cases_per_million                  107274 non-null  float64
 12  new_cases_smoothed_per_million         106264 non-null  float64
 13  total_deaths_per_million               96743 non-null   float64
 14  new_deaths_per_million                 96898 non-null   float64
 15  new_deaths_smoothed_per_million        106264 non-null  float64
 16  reproduction_rate                      90551 non-null   float64
 17  icu_patients                           12289 non-null   float64
 18  icu_patients_per_million               12289 non-null   float64
 19  hosp_patients                          14703 non-null   float64
 20  hosp_patients_per_million              14703 non-null   float64
 21  weekly_icu_admissions                  1111 non-null    float64
 22  weekly_icu_admissions_per_million      1111 non-null    float64
 23  weekly_hosp_admissions                 1966 non-null    float64
 24  weekly_hosp_admissions_per_million     1966 non-null    float64
 25  new_tests                              49431 non-null   float64
 26  total_tests                            49210 non-null   float64
 27  total_tests_per_thousand               49210 non-null   float64
 28  new_tests_per_thousand                 49431 non-null   float64
 29  new_tests_smoothed                     58246 non-null   float64
 30  new_tests_smoothed_per_thousand        58246 non-null   float64
 31  positive_rate                          54481 non-null   float64
 32  tests_per_case                         53865 non-null   float64
 33  tests_units                            60083 non-null   object 
 34  total_vaccinations                     23382 non-null   float64
 35  people_vaccinated                      22393 non-null   float64
 36  people_fully_vaccinated                19448 non-null   float64
 37  total_boosters                         381 non-null     float64
 38  new_vaccinations                       19407 non-null   float64
 39  new_vaccinations_smoothed              41293 non-null   float64
 40  total_vaccinations_per_hundred         23382 non-null   float64
 41  people_vaccinated_per_hundred          22393 non-null   float64
 42  people_fully_vaccinated_per_hundred    19448 non-null   float64
 43  total_boosters_per_hundred             381 non-null     float64
 44  new_vaccinations_smoothed_per_million  41293 non-null   float64
 45  stringency_index                       93794 non-null   float64
 46  population                             112080 non-null  float64
 47  population_density                     104581 non-null  float64
 48  median_age                             100213 non-null  float64
 49  aged_65_older                          99095 non-null   float64
 50  aged_70_older                          99662 non-null   float64
 51  gdp_per_capita                         100654 non-null  float64
 52  extreme_poverty                        67675 non-null   float64
 53  cardiovasc_death_rate                  100536 non-null  float64
 54  diabetes_prevalence                    103384 non-null  float64
 55  female_smokers                         78518 non-null   float64
 56  male_smokers                           77374 non-null   float64
 57  handwashing_facilities                 50502 non-null   float64
 58  hospital_beds_per_thousand             91442 non-null   float64
 59  life_expectancy                        107075 non-null  float64
 60  human_development_index                100632 non-null  float64
 61  excess_mortality                       3922 non-null    float64
dtypes: float64(57), object(5)
memory usage: 53.4+ MB


#Getting list of ISO codes which are unusual

for r in df.iso_code.unique():
    if 'OWID' in r:
        print(r)

OWID_AFR
OWID_ASI
OWID_EUR
OWID_EUN
OWID_INT
OWID_KOS
OWID_NAM
OWID_CYN
OWID_OCE
OWID_SAM
OWID_WRL


#Checking rows associated with the ISO code 'OWID_CYN'

df[df['iso_code'] == 'OWID_CYN']


#Dropping unrequired columns, changing format to datetime, renaming column, renaming ISO code for Kosovo, dropping
#Northern Cyprus as it's all missing data (and not a recognized nation)

cols_to_drop = [21,22,23,24,33]
df.drop(df.columns[cols_to_drop], inplace=True, axis=1)
df['date'] = pd.to_datetime(df['date'], format=('%Y-%m-%d'))
df.rename(columns = {'location':'country'},inplace=True)
df.replace('OWID_KOS','XKX', inplace=True)
df.drop(df[df['iso_code']=='OWID_CYN'].index, inplace=True)


df.shape

(112650, 57)


#Importing the data with recovery stats

if STATIC:
    recovered_df = pd.read_csv('static_raw_recovered_df.csv')
else:
    recovered_df = pd.read_csv('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/\
csse_covid_19_time_series/time_series_covid19_recovered_global.csv')
    
recovered_df.head()


# #Saving to local

# recovered_df.to_csv('static_raw_recovered_df.csv', index=False)


#Dropping all the columns from 5th August onwards

recovered_df = recovered_df.loc[:,:'8/4/21']


#Getting list of countries with multiple occurences (due to Province/State) to add them up into a single row

recovered_countries =  list(recovered_df['Country/Region'])
counter_a = Counter(recovered_countries)

country_with_multiple_instances = []

for country_name in counter_a:
    if counter_a[country_name] > 1:
        print(country_name)
        country_with_multiple_instances.append(country_name)

Australia
China
Denmark
France
Netherlands
New Zealand
United Kingdom


#Appending country-wise dataframes to a list to concat into a single dataframe later

date_cols = recovered_df.columns[4:]
full_recovered_df = []

for country in set(recovered_countries):
    rec_df = recovered_df[recovered_df['Country/Region'] == country][date_cols].T
    if country in country_with_multiple_instances:
        rec_df['temp'] = rec_df.sum(axis=1)
        rec_df.drop(rec_df.iloc[:,:-1], axis=1, inplace=True)
    lat = recovered_df[recovered_df['Country/Region'] == country]['Lat'].values[0]
    long = recovered_df[recovered_df['Country/Region'] == country]['Long'].values[0]
    rec_df.columns = ['Recovered']
    rec_df['Country'] = country
    rec_df['Latitude'] = lat
    rec_df['Longitude'] = long
    full_recovered_df.append(rec_df)


#Transforming list of dataframes into a single dataframe and cleaning it up

recovered_df = pd.concat(full_recovered_df)
recovered_df.reset_index(inplace=True)
recovered_df.rename(columns = {'index':'Date'},inplace=True)
recovered_df['Date'] = recovered_df['Date'].apply(lambda x:x[:-2] + '20' + x[-2:])
recovered_df['Date'] = pd.to_datetime(recovered_df['Date'],format=('%m/%d/%Y'))
recovered_df.sort_values(['Country','Date'],inplace=True)
recovered_df.head()


#Setting coordinates for a few countries to np.nan due to multiple regions for them in dataframe

for country in country_with_multiple_instances:
    recovered_df.loc[recovered_df['Country'] == country, 'Latitude'] = np.nan
    recovered_df.loc[recovered_df['Country'] == country, 'Longitude'] = np.nan


#Getting list of countries exclusive to each dataframe by using XOR

set(recovered_countries) ^ set(df['country'].unique())

{'Africa',
 'Anguilla',
 'Aruba',
 'Asia',
 'Bermuda',
 'Bonaire Sint Eustatius and Saba',
 'British Virgin Islands',
 'Burma',
 'Cabo Verde',
 'Cape Verde',
 'Cayman Islands',
 'Congo',
 'Congo (Brazzaville)',
 'Congo (Kinshasa)',
 'Cook Islands',
 'Curacao',
 'Democratic Republic of Congo',
 'Diamond Princess',
 'Europe',
 'European Union',
 'Faeroe Islands',
 'Falkland Islands',
 'French Polynesia',
 'Gibraltar',
 'Greenland',
 'Guernsey',
 'Holy See',
 'Hong Kong',
 'International',
 'Isle of Man',
 'Jersey',
 'Korea, South',
 'MS Zaandam',
 'Macao',
 'Micronesia',
 'Micronesia (country)',
 'Montserrat',
 'Myanmar',
 'Nauru',
 'New Caledonia',
 'Niue',
 'North America',
 'Oceania',
 'Palestine',
 'Pitcairn',
 'Saint Helena',
 'Sint Maarten (Dutch part)',
 'South America',
 'South Korea',
 'Summer Olympics 2020',
 'Taiwan',
 'Taiwan*',
 'Timor',
 'Timor-Leste',
 'Tokelau',
 'Tonga',
 'Turkmenistan',
 'Turks and Caicos Islands',
 'Tuvalu',
 'US',
 'United States',
 'Vatican',
 'Wallis and Futuna',
 'West Bank and Gaza',
 'World'}


#Renaming country names to the same standard

recovered_df.replace({'Burma':'Myanmar', 'Cabo Verde':'Cape Verde', 'Congo (Brazzaville)':'Congo',
                            'Congo (Kinshasa)':'Democratic Republic of Congo',
                            'Korea, South':'South Korea', 'Taiwan*':'Taiwan',
                            'Timor-Leste':'Timor', 'US':'United States'}, inplace=True)


#Renaming columns and setting indexes to prepare for merge

recovered_df.rename(columns={'Date':'date','Country':'country','Recovered':'total_recovered',
                            'Latitude':'latitude', 'Longitude':'longitude'}, inplace=True)
recovered_df.set_index(['date','country'], inplace=True)
df.set_index(['date','country'], inplace=True)


#Merging (df) and (recovered_df)

df = df.join(recovered_df)


df.shape

(112650, 58)


#Importing government response data

if STATIC:
    df_gov = pd.read_csv('static_raw_df_gov.csv')
else:
    df_gov = pd.read_csv('https://raw.githubusercontent.com/OxCGRT/covid-policy-tracker/master/data/OxCGRT_latest.csv')
    
df_gov.head()


# #Saving to local

# df_gov.to_csv('static_raw_df_gov.csv', index=False)


df_gov.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 189696 entries, 0 to 189695
Data columns (total 51 columns):
 #   Column                                 Non-Null Count   Dtype  
---  ------                                 --------------   -----  
 0   CountryName                            189696 non-null  object 
 1   CountryCode                            189696 non-null  object 
 2   RegionName                             76608 non-null   object 
 3   RegionCode                             76608 non-null   object 
 4   Jurisdiction                           189696 non-null  object 
 5   Date                                   189696 non-null  int64  
 6   C1_School closing                      183573 non-null  float64
 7   C1_Flag                                150484 non-null  float64
 8   C2_Workplace closing                   183491 non-null  float64
 9   C2_Flag                                143032 non-null  float64
 10  C3_Cancel public events                183568 non-null  float64
 11  C3_Flag                                152917 non-null  float64
 12  C4_Restrictions on gatherings          183546 non-null  float64
 13  C4_Flag                                138941 non-null  float64
 14  C5_Close public transport              183596 non-null  float64
 15  C5_Flag                                86963 non-null   float64
 16  C6_Stay at home requirements           183441 non-null  float64
 17  C6_Flag                                119288 non-null  float64
 18  C7_Restrictions on internal movement   183500 non-null  float64
 19  C7_Flag                                115363 non-null  float64
 20  C8_International travel controls       181532 non-null  float64
 21  E1_Income support                      183300 non-null  float64
 22  E1_Flag                                116497 non-null  float64
 23  E2_Debt/contract relief                182621 non-null  float64
 24  E3_Fiscal measures                     146636 non-null  float64
 25  E4_International support               146532 non-null  float64
 26  H1_Public information campaigns        183147 non-null  float64
 27  H1_Flag                                170471 non-null  float64
 28  H2_Testing policy                      182989 non-null  float64
 29  H3_Contact tracing                     182564 non-null  float64
 30  H4_Emergency investment in healthcare  146359 non-null  float64
 31  H5_Investment in vaccines              149028 non-null  float64
 32  H6_Facial Coverings                    182863 non-null  float64
 33  H6_Flag                                146230 non-null  float64
 34  H7_Vaccination policy                  166174 non-null  float64
 35  H7_Flag                                59800 non-null   float64
 36  H8_Protection of elderly people        182073 non-null  float64
 37  H8_Flag                                134596 non-null  float64
 38  M1_Wildcard                            0 non-null       float64
 39  ConfirmedCases                         179406 non-null  float64
 40  ConfirmedDeaths                        179329 non-null  float64
 41  StringencyIndex                        183157 non-null  float64
 42  StringencyIndexForDisplay              184128 non-null  float64
 43  StringencyLegacyIndex                  183258 non-null  float64
 44  StringencyLegacyIndexForDisplay        184227 non-null  float64
 45  GovernmentResponseIndex                182522 non-null  float64
 46  GovernmentResponseIndexForDisplay      183500 non-null  float64
 47  ContainmentHealthIndex                 182676 non-null  float64
 48  ContainmentHealthIndexForDisplay       183647 non-null  float64
 49  EconomicSupportIndex                   182341 non-null  float64
 50  EconomicSupportIndexForDisplay         183320 non-null  float64
dtypes: float64(45), int64(1), object(5)
memory usage: 73.8+ MB


#Dropping columns and rows not needed, renaming other columns, changing date format

cols_to_drop2 = [0,2,3,4,7,9,11,13,15,17,19,22,27,33,35,37,38,39,40,41,42,43,44,46,48,50]
df_gov.drop(df_gov[df_gov['Jurisdiction'] == 'STATE_TOTAL'].index, inplace=True)
df_gov.drop(df_gov.columns[cols_to_drop2],inplace=True, axis=1)
modifying_col_names = df_gov.columns.str.split('_')
new_col_names = [col_name[0] if len(col_name) is 1 else col_name[1] for col_name in modifying_col_names]
df_gov.columns = new_col_names
df_gov['Date'] = pd.to_datetime(df_gov['Date'],format=('%Y%m%d'))


df_gov.head()


#Renaming columns and setting index to prepare for merge

df_gov.rename(columns={'Date':'date','CountryCode':'iso_code'},inplace=True)
df_gov.set_index(['date','iso_code'],inplace=True)
df.reset_index(inplace=True)
df.set_index(['date','iso_code'],inplace=True)


#Merging (df) and (df_gov)

df = df.join(df_gov)
df.reset_index(inplace=True)


df.shape

(112650, 83)


#Importing google mobility trends data (700+ MB data)

if STATIC:
    google_df = pd.read_csv('static_raw_google_df.csv')
else:
    google_df = pd.read_csv('https://www.gstatic.com/covid19/mobility/Global_Mobility_Report.csv')
    
google_df.head()


# #Saving to local

# google_df.to_csv('static_raw_google_df.csv', index=False)


#Dropping data for sub-regions inside the countries, keeping only the whole country data

google_df = google_df[pd.isnull(google_df['sub_region_1'])]
google_df = google_df[pd.isnull(google_df['metro_area'])]


google_df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 75098 entries, 0 to 6652836
Data columns (total 15 columns):
 #   Column                                              Non-Null Count  Dtype  
---  ------                                              --------------  -----  
 0   country_region_code                                 74540 non-null  object 
 1   country_region                                      75098 non-null  object 
 2   sub_region_1                                        0 non-null      object 
 3   sub_region_2                                        0 non-null      object 
 4   metro_area                                          0 non-null      object 
 5   iso_3166_2_code                                     0 non-null      object 
 6   census_fips_code                                    0 non-null      float64
 7   place_id                                            75098 non-null  object 
 8   date                                                75098 non-null  object 
 9   retail_and_recreation_percent_change_from_baseline  73730 non-null  float64
 10  grocery_and_pharmacy_percent_change_from_baseline   73664 non-null  float64
 11  parks_percent_change_from_baseline                  73642 non-null  float64
 12  transit_stations_percent_change_from_baseline       74073 non-null  float64
 13  workplaces_percent_change_from_baseline             74778 non-null  float64
 14  residential_percent_change_from_baseline            73338 non-null  float64
dtypes: float64(7), object(8)
memory usage: 9.2+ MB


#Dropping columns not needed, renaming other columns

cols_to_drop3 = [0,2,3,4,5,6,7]
google_df.drop(google_df.columns[cols_to_drop3], axis=1, inplace=True)
google_df.rename(columns = {'country_region':'country'})
google_df.columns = ['country','date'] + [("gm_"+col_name[:-29]) for col_name in google_df.columns[2:]]


#Getting country names in google_df but not in original df

for country in google_df['country'].unique():
    if country not in df['country'].unique():
        print(country)

The Bahamas
Côte d'Ivoire
Myanmar (Burma)
Puerto Rico
Réunion


#Renaming some countries, dropping others (overseas regions) as they don't exist in the main dataframe

google_df.replace({'The Bahamas':'Bahamas', "Côte d'Ivoire":"Cote d'Ivoire",
                  'Myanmar (Burma)':'Myanmar'},inplace=True)


#Formatting google_df and merging it with original df

google_df['date'] = pd.to_datetime(google_df['date'], format=('%Y-%m-%d'))
google_df.set_index(['date','country'],inplace=True)

df.set_index(['date','country'], inplace=True)
df = df.join(google_df)
df.reset_index(inplace=True)


df.shape

(112650, 89)


#Dropping rows 'International', 'European Union' and 'Macao' (all NaN values) in countries. And renaming Micronesia
#to proper name

df.drop(df[(df['iso_code'] == 'OWID_INT') | (df['iso_code'] == 'OWID_EUN')].index, inplace=True)
df.drop(df[df['country'] == 'Macao'].index, inplace=True)
df.replace({'Micronesia (country)':'Micronesia'}, inplace=True)


#Checking for missing coordinate values

df['latitude'].isna().value_counts()

False    92689
True     18609
Name: latitude, dtype: int64


#Adding all coordinates to dictionary, np.nan for missing coordinates

lat_longs = {}
arbitrary_mid_date = '2020-05-01'

for country in df['country'].unique():
    try:
        lat = df[(df['country'] == country) & (df['date'] == arbitrary_mid_date)].latitude.values[0]
        long = df[(df['country'] == country) & (df['date'] == arbitrary_mid_date)].longitude.values[0]
    except IndexError:
        lat_longs[country] = {'lat':np.nan, 'long':np.nan}
        continue
    lat_longs[country] = {'lat':lat, 'long':long}
    
for i,j in lat_longs.items():
    if np.isnan(j['lat']):
        print(i,j)

Africa {'lat': nan, 'long': nan}
Anguilla {'lat': nan, 'long': nan}
Aruba {'lat': nan, 'long': nan}
Asia {'lat': nan, 'long': nan}
Australia {'lat': nan, 'long': nan}
Bermuda {'lat': nan, 'long': nan}
Bonaire Sint Eustatius and Saba {'lat': nan, 'long': nan}
British Virgin Islands {'lat': nan, 'long': nan}
Cayman Islands {'lat': nan, 'long': nan}
China {'lat': nan, 'long': nan}
Cook Islands {'lat': nan, 'long': nan}
Curacao {'lat': nan, 'long': nan}
Denmark {'lat': nan, 'long': nan}
Europe {'lat': nan, 'long': nan}
Faeroe Islands {'lat': nan, 'long': nan}
Falkland Islands {'lat': nan, 'long': nan}
France {'lat': nan, 'long': nan}
French Polynesia {'lat': nan, 'long': nan}
Gibraltar {'lat': nan, 'long': nan}
Greenland {'lat': nan, 'long': nan}
Guernsey {'lat': nan, 'long': nan}
Hong Kong {'lat': nan, 'long': nan}
Isle of Man {'lat': nan, 'long': nan}
Jersey {'lat': nan, 'long': nan}
Kiribati {'lat': nan, 'long': nan}
Lesotho {'lat': nan, 'long': nan}
Marshall Islands {'lat': nan, 'long': nan}
Micronesia {'lat': nan, 'long': nan}
Montserrat {'lat': nan, 'long': nan}
Nauru {'lat': nan, 'long': nan}
Netherlands {'lat': nan, 'long': nan}
New Caledonia {'lat': nan, 'long': nan}
New Zealand {'lat': nan, 'long': nan}
Niue {'lat': nan, 'long': nan}
North America {'lat': nan, 'long': nan}
Oceania {'lat': nan, 'long': nan}
Palau {'lat': nan, 'long': nan}
Palestine {'lat': nan, 'long': nan}
Pitcairn {'lat': nan, 'long': nan}
Saint Helena {'lat': nan, 'long': nan}
Samoa {'lat': nan, 'long': nan}
Sint Maarten (Dutch part) {'lat': nan, 'long': nan}
Solomon Islands {'lat': nan, 'long': nan}
South America {'lat': nan, 'long': nan}
Tokelau {'lat': nan, 'long': nan}
Tonga {'lat': nan, 'long': nan}
Turkmenistan {'lat': nan, 'long': nan}
Turks and Caicos Islands {'lat': nan, 'long': nan}
Tuvalu {'lat': nan, 'long': nan}
United Kingdom {'lat': nan, 'long': nan}
Vanuatu {'lat': nan, 'long': nan}
Vatican {'lat': nan, 'long': nan}
Wallis and Futuna {'lat': nan, 'long': nan}
World {'lat': nan, 'long': nan}


#Getting coordinates for countries with np.nan values

geolocator = Nominatim(user_agent='abc')
for country,coords in lat_longs.items():
    if not np.isnan(coords['lat']):
        continue
    try:
        location = geolocator.geocode(country)
        lat = location.latitude
        long = location.longitude
        lat_longs[country] = {'lat':lat, 'long':long}
    except:
        print(country)


#Deleting 'World' and continents from coordinate dict

keys_to_remove = ['World','Europe','North America','Asia','South America','Oceania','Africa']

for key in keys_to_remove:
    lat_longs.pop(key)


#Filling latitudes and longitudes for all the countries

for country in df['country'].unique():
    if country in keys_to_remove:
        continue
    lat = lat_longs[country]['lat']
    long = lat_longs[country]['long']
    df.loc[df['country']==country, ['latitude','longitude']] = (lat, long)


df.shape

(111298, 89)


#Importing a dataset which contains literacy rates for all countries

df_literacy = pd.read_csv('https://raw.githubusercontent.com/vyaduvanshi/helper-files/master/\
literacy-rates-most-recent.csv')

df_literacy.head()


#Changing ISO code of Kosovo to project norm, and dropping some unrequired rows

df_literacy.replace({'OWID_KOS':'XKX'}, inplace=True)
df_literacy.drop(df_literacy[df_literacy['Code']=='OWID_WRL'].index, inplace=True)
df_literacy.dropna(axis=0, subset=['Code'], inplace=True)


#Cleaning literacy_df and merging it to the main dataframe

df_literacy.rename(columns={'Code':'iso_code',
                            'Literacy rates (World Bank, CIA World Factbook, and other sources)':'literacy_rate'},
                   inplace=True)
df_literacy = pd.pivot_table(df_literacy, index='iso_code', aggfunc='last')
df_literacy.drop(['Entity','Year'], axis=1, inplace=True)

df.set_index('iso_code',inplace=True)
df = df.join(df_literacy)
df.reset_index(inplace=True)


df.shape

(111298, 90)


#Importing democracy index (di) country-wise

di_df = pd.read_csv('https://raw.githubusercontent.com/vyaduvanshi/helper-files/master/democracy_index.csv')
di_df.head()


#Renaming, formatting and dropping a column from di_df

di_df.dropna(inplace=True)
di_df.rename(columns={'iso':'iso_code'}, inplace=True)
di_df.iso_code = di_df.iso_code.str.upper()
di_df.drop(['country'], axis=1, inplace=True)


#Dropping rows and columns not needed and merging di_df with main dataframe

# di_df.drop(di_df[(di_df['iso_code']=='PRK') | (di_df['iso_code']=='TKM')].index, inplace=True)
di_df = di_df.pivot_table(di_df, index='iso_code', aggfunc='last')
del di_df['year']

df.set_index('iso_code', inplace=True)
df = df.join(di_df)
df.reset_index(inplace=True)


df.shape

(111298, 91)


df.head()


#Setting recovered values for US to NaN because of null records after 13th Dec

us_recovered_df = df[df['country'] == 'United States']
df.loc[us_recovered_df.index, 'total_recovered'] = np.nan


#Creating two separate dataframes, one for continents and one for world

continent_list = ['Africa','Asia','Europe','Oceania','North America','South America']
continent_df = []

for continent in continent_list:
    temp_df = df[df['country'] == continent]
    continent_df.append(temp_df)

continent_df = pd.concat(continent_df)
world_df = df[df['country'] == 'World']


#Dropping 'World' and continents in 'country' column as they are already separated into their own dataframes

for key in keys_to_remove:
    df.drop(df[df['country'] == key].index, inplace=True)


#Setting the date values in `world_df` and `continent_df` same as main `df`

world_df = world_df.set_index('date').reindex(np.sort(df.date.unique())).reset_index()
continent_df = continent_df.set_index('date').groupby('country').apply(lambda x: x.reindex(pd.date_range(min(df.date),
                                                                                                         max(df.date),
                                                                                                         freq = 'D')))\
.drop('country', axis=1).reset_index()
continent_df.rename(columns = {continent_df.columns[1]:'date'}, inplace=True)


#Function to calculate moving average for a grouped dataframe
def calc_MA_on_group(x):
    fill_val = x.rolling(window=7).mean()
    return pd.Series(fill_val, index=x.index)


#Function to create new features in a dataframe
def create_features(df, group=None):
    df['percent_pop_affected'] = df['total_cases']/df['population'] * 100
    df['percent_pop_tested'] = df['total_tests']/df['population'] * 100
    df['percent_pop_vaccinated'] = df['total_vaccinations']/df['population'] * 100
    df['herd_immunity_threshold'] = 0.67 * df['population']
    df['death_percent'] = df['total_deaths']/df['total_cases'] * 100
    df['recovered_percent'] = df['total_recovered']/df['total_cases'] * 100
    df['total_cases_per_hundred'] = (df['total_cases'] * 100)/df['population']
    df['total_deaths_per_hundred'] = (df['total_deaths'] * 100)/df['population']
    df['total_recovered_per_hundred'] = (df['total_recovered'] * 100)/df['population']
    df['total_tests_per_hundred'] = (df['total_tests'] * 100)/df['population']
    
    if group == 'world':
        df['population'] = df['population'].bfill()
        df['herd_immunity_threshold'] = 0.67 * df['population']
        df['positive_rate'] = (df['new_cases']/df['new_tests']).rolling(window=7).mean()
        df['tests_per_case'] =  (df['new_tests']/df['new_cases']).rolling(window=7).mean()
        df['new_tests_smoothed'] = df['new_tests'].rolling(window=7).mean()
        df['new_recovered_smoothed'] = df['new_recovered'].rolling(window=7).mean()
        df['percent_increase_cases'] = df['new_cases'].pct_change() * 100
        
    if group == 'all':
        df['active_cases'] = df['total_cases'] - df['total_recovered'] - df['total_deaths']
        df['new_recovered'] = df.groupby('country')['total_recovered'].diff(1)
        df['total_tests'] = df.groupby('country')['total_tests'].ffill()
        df['total_recovered'] = df.groupby('country')['total_recovered'].ffill()
        df['total_cases'] = df.groupby('country')['total_cases'].ffill()
        df['total_vaccinations'] = df.groupby('country')['total_vaccinations'].ffill()
        df['total_deaths'] = df.groupby('country')['total_deaths'].ffill()
        
    if group == 'continent' or group == 'all':
        df['new_recovered_smoothed'] = df.groupby('country')['new_recovered'].apply(calc_MA_on_group)
        df['percent_increase_cases'] = df.groupby('country')['new_cases'].pct_change() * 100
        
    if group == 'continent':
        df['new_tests_smoothed'] = df.groupby('country')['new_tests'].apply(calc_MA_on_group)
        df['positive_rate_smoothed'] = df['new_cases']/df['new_tests']
        df['positive_rate_smoothed'] = df.groupby('country')['positive_rate_smoothed'].apply(calc_MA_on_group)
        df['tests_per_case_smoothed'] = df['new_tests']/df['new_cases']
        df['tests_per_case_smoothed'] = df.groupby('country')['tests_per_case_smoothed'].apply(calc_MA_on_group)


#Creating new features in main dataframe

create_features(df, 'all')


#Creating temporary dataframes for World and continents to calculate new features for main 'world_df' & 'continent_df'

temp_world_df = pd.pivot_table(df, index='date', aggfunc=np.sum).reset_index()
temp_continent_df = pd.pivot_table(df, index=['continent','date'], aggfunc=np.sum).reset_index()

temp_world_df = temp_world_df[['date','new_tests','icu_patients','hosp_patients','total_tests','total_recovered',
                               'new_recovered','active_cases']]
temp_continent_df = temp_continent_df[['date','continent','total_recovered','new_recovered','total_tests',
                                      'new_tests','active_cases']]


#Dropping all empty or unrequired columns from 'world_df' and 'continent_df'

world_df = world_df.dropna(axis=1, how='all').drop(columns = ['iso_code','country'])
continent_df = continent_df.dropna(axis=1, how='all').drop(columns = ['iso_code'])


#Joining `world_df` and `continent_df` with `temp_world_df` and `temp_continent_df`

world_df = world_df.set_index('date').join(temp_world_df.set_index('date'))
continent_df = continent_df.join(temp_continent_df.set_index(['date','continent']), on=['date','country'])


#Creating new features from function for both dataframes and renaming a column in `continent_df`

world_df.reset_index(inplace=True)
create_features(world_df, 'world')
create_features(continent_df, 'continent')

continent_df.rename(columns={'country':'continent'}, inplace=True)


#Creating two dataframes grouped into intervals of 25 days each for plotting

continent_df_25d = continent_df.groupby(['continent', pd.Grouper(key='date', freq='25d')]).last().reset_index()
df_25d = df.groupby(['continent','country', pd.Grouper(key='date', freq='25d')]).last().reset_index()


# #Saving three stable and static final dataframes (df, world_df, continent_df) to disk

# df.to_csv('final_df.csv', index=False)
# world_df.to_csv('final_world_df.csv',index=False)
# continent_df.to_csv('final_continent_df.csv', index=False)


#Importing from local. Necessary if you jumped directly to EDA

df = pd.read_csv('final_df.csv')
world_df = pd.read_csv('final_world_df.csv')
continent_df = pd.read_csv('final_continent_df.csv')

df['date'] = pd.to_datetime(df['date'], format=('%Y-%m-%d'))
world_df['date'] = pd.to_datetime(world_df['date'], format=('%Y-%m-%d'))
continent_df['date'] = pd.to_datetime(continent_df['date'], format=('%Y-%m-%d'))


df.info(verbose=True)

<class 'pandas.core.frame.DataFrame'>
Int64Index: 107253 entries, 0 to 111297
Data columns (total 105 columns):
 #    Column                                 Dtype         
---   ------                                 -----         
 0    iso_code                               object        
 1    date                                   datetime64[ns]
 2    country                                object        
 3    continent                              object        
 4    total_cases                            float64       
 5    new_cases                              float64       
 6    new_cases_smoothed                     float64       
 7    total_deaths                           float64       
 8    new_deaths                             float64       
 9    new_deaths_smoothed                    float64       
 10   total_cases_per_million                float64       
 11   new_cases_per_million                  float64       
 12   new_cases_smoothed_per_million         float64       
 13   total_deaths_per_million               float64       
 14   new_deaths_per_million                 float64       
 15   new_deaths_smoothed_per_million        float64       
 16   reproduction_rate                      float64       
 17   icu_patients                           float64       
 18   icu_patients_per_million               float64       
 19   hosp_patients                          float64       
 20   hosp_patients_per_million              float64       
 21   new_tests                              float64       
 22   total_tests                            float64       
 23   total_tests_per_thousand               float64       
 24   new_tests_per_thousand                 float64       
 25   new_tests_smoothed                     float64       
 26   new_tests_smoothed_per_thousand        float64       
 27   positive_rate                          float64       
 28   tests_per_case                         float64       
 29   total_vaccinations                     float64       
 30   people_vaccinated                      float64       
 31   people_fully_vaccinated                float64       
 32   total_boosters                         float64       
 33   new_vaccinations                       float64       
 34   new_vaccinations_smoothed              float64       
 35   total_vaccinations_per_hundred         float64       
 36   people_vaccinated_per_hundred          float64       
 37   people_fully_vaccinated_per_hundred    float64       
 38   total_boosters_per_hundred             float64       
 39   new_vaccinations_smoothed_per_million  float64       
 40   stringency_index                       float64       
 41   population                             float64       
 42   population_density                     float64       
 43   median_age                             float64       
 44   aged_65_older                          float64       
 45   aged_70_older                          float64       
 46   gdp_per_capita                         float64       
 47   extreme_poverty                        float64       
 48   cardiovasc_death_rate                  float64       
 49   diabetes_prevalence                    float64       
 50   female_smokers                         float64       
 51   male_smokers                           float64       
 52   handwashing_facilities                 float64       
 53   hospital_beds_per_thousand             float64       
 54   life_expectancy                        float64       
 55   human_development_index                float64       
 56   excess_mortality                       float64       
 57   total_recovered                        float64       
 58   latitude                               float64       
 59   longitude                              float64       
 60   School closing                         float64       
 61   Workplace closing                      float64       
 62   Cancel public events                   float64       
 63   Restrictions on gatherings             float64       
 64   Close public transport                 float64       
 65   Stay at home requirements              float64       
 66   Restrictions on internal movement      float64       
 67   International travel controls          float64       
 68   Income support                         float64       
 69   Debt/contract relief                   float64       
 70   Fiscal measures                        float64       
 71   International support                  float64       
 72   Public information campaigns           float64       
 73   Testing policy                         float64       
 74   Contact tracing                        float64       
 75   Emergency investment in healthcare     float64       
 76   Investment in vaccines                 float64       
 77   Facial Coverings                       float64       
 78   Vaccination policy                     float64       
 79   Protection of elderly people           float64       
 80   GovernmentResponseIndex                float64       
 81   ContainmentHealthIndex                 float64       
 82   EconomicSupportIndex                   float64       
 83   gm_retail_and_recreation               float64       
 84   gm_grocery_and_pharmacy                float64       
 85   gm_parks                               float64       
 86   gm_transit_stations                    float64       
 87   gm_workplaces                          float64       
 88   gm_residential                         float64       
 89   literacy_rate                          float64       
 90   Democracy index                        float64       
 91   percent_pop_affected                   float64       
 92   percent_pop_tested                     float64       
 93   percent_pop_vaccinated                 float64       
 94   herd_immunity_threshold                float64       
 95   death_percent                          float64       
 96   recovered_percent                      float64       
 97   total_cases_per_hundred                float64       
 98   total_deaths_per_hundred               float64       
 99   total_recovered_per_hundred            float64       
 100  total_tests_per_hundred                float64       
 101  active_cases                           float64       
 102  new_recovered                          float64       
 103  new_recovered_smoothed                 float64       
 104  percent_increase_cases                 float64       
dtypes: datetime64[ns](1), float64(101), object(3)
memory usage: 86.7+ MB


#Custom theme for plotly graphs

my_theme = go.layout.Template({
    'data': {'bar': [{'error_x': {'color': '#2a3f5f'},
                      'error_y': {'color': '#2a3f5f'},
                      'marker': {'line': {'color': '#E5ECF6', 'width': 0.5}},
                      'type': 'bar'}],
             'barpolar': [{'marker': {'line': {'color': '#E5ECF6', 'width': 0.5}}, 'type': 'barpolar'}],
             'carpet': [{'aaxis': {'endlinecolor': '#2a3f5f',
                                   'gridcolor': 'white',
                                   'linecolor': 'white',
                                   'minorgridcolor': 'white',
                                   'startlinecolor': '#2a3f5f'},
                         'baxis': {'endlinecolor': '#2a3f5f',
                                   'gridcolor': 'white',
                                   'linecolor': 'white',
                                   'minorgridcolor': 'white',
                                   'startlinecolor': '#2a3f5f'},
                         'type': 'carpet'}],
             'choropleth': [{'colorbar': {'outlinewidth': 0, 'ticks': ''}, 'type': 'choropleth'}],
             'contour': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                          'colorscale': [[0.0, '#0d0887'], [0.1111111111111111,
                                         '#46039f'], [0.2222222222222222,
                                         '#7201a8'], [0.3333333333333333,
                                         '#9c179e'], [0.4444444444444444,
                                         '#bd3786'], [0.5555555555555556,
                                         '#d8576b'], [0.6666666666666666,
                                         '#ed7953'], [0.7777777777777778,
                                         '#fb9f3a'], [0.8888888888888888,
                                         '#fdca26'], [1.0, '#f0f921']],
                          'type': 'contour'}],
             'contourcarpet': [{'colorbar': {'outlinewidth': 0, 'ticks': ''}, 'type': 'contourcarpet'}],
             'heatmap': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                          'colorscale': [[0.0, '#0d0887'], [0.1111111111111111,
                                         '#46039f'], [0.2222222222222222,
                                         '#7201a8'], [0.3333333333333333,
                                         '#9c179e'], [0.4444444444444444,
                                         '#bd3786'], [0.5555555555555556,
                                         '#d8576b'], [0.6666666666666666,
                                         '#ed7953'], [0.7777777777777778,
                                         '#fb9f3a'], [0.8888888888888888,
                                         '#fdca26'], [1.0, '#f0f921']],
                          'type': 'heatmap'}],
             'heatmapgl': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                            'colorscale': [[0.0, '#0d0887'], [0.1111111111111111,
                                           '#46039f'], [0.2222222222222222,
                                           '#7201a8'], [0.3333333333333333,
                                           '#9c179e'], [0.4444444444444444,
                                           '#bd3786'], [0.5555555555555556,
                                           '#d8576b'], [0.6666666666666666,
                                           '#ed7953'], [0.7777777777777778,
                                           '#fb9f3a'], [0.8888888888888888,
                                           '#fdca26'], [1.0, '#f0f921']],
                            'type': 'heatmapgl'}],
             'histogram': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'histogram'}],
             'histogram2d': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                              'colorscale': [[0.0, '#0d0887'],
                                             [0.1111111111111111, '#46039f'],
                                             [0.2222222222222222, '#7201a8'],
                                             [0.3333333333333333, '#9c179e'],
                                             [0.4444444444444444, '#bd3786'],
                                             [0.5555555555555556, '#d8576b'],
                                             [0.6666666666666666, '#ed7953'],
                                             [0.7777777777777778, '#fb9f3a'],
                                             [0.8888888888888888, '#fdca26'], [1.0,
                                             '#f0f921']],
                              'type': 'histogram2d'}],
             'histogram2dcontour': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                                     'colorscale': [[0.0, '#0d0887'],
                                                    [0.1111111111111111,
                                                    '#46039f'],
                                                    [0.2222222222222222,
                                                    '#7201a8'],
                                                    [0.3333333333333333,
                                                    '#9c179e'],
                                                    [0.4444444444444444,
                                                    '#bd3786'],
                                                    [0.5555555555555556,
                                                    '#d8576b'],
                                                    [0.6666666666666666,
                                                    '#ed7953'],
                                                    [0.7777777777777778,
                                                    '#fb9f3a'],
                                                    [0.8888888888888888,
                                                    '#fdca26'], [1.0, '#f0f921']],
                                     'type': 'histogram2dcontour'}],
             'mesh3d': [{'colorbar': {'outlinewidth': 0, 'ticks': ''}, 'type': 'mesh3d'}],
             'parcoords': [{'line': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'parcoords'}],
             'pie': [{'automargin': True, 'type': 'pie'}],
             'scatter': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scatter',
                         'line':{'width':2.5}}],
             'scatter3d': [{'line': {'colorbar': {'outlinewidth': 0, 'ticks': ''}},
                            'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}},
                            'type': 'scatter3d'}],
             'scattercarpet': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scattercarpet'}],
             'scattergeo': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scattergeo'}],
             'scattergl': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scattergl'}],
             'scattermapbox': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scattermapbox'}],
             'scatterpolar': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scatterpolar'}],
             'scatterpolargl': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scatterpolargl'}],
             'scatterternary': [{'marker': {'colorbar': {'outlinewidth': 0, 'ticks': ''}}, 'type': 'scatterternary'}],
             'surface': [{'colorbar': {'outlinewidth': 0, 'ticks': ''},
                          'colorscale': [[0.0, '#0d0887'], [0.1111111111111111,
                                         '#46039f'], [0.2222222222222222,
                                         '#7201a8'], [0.3333333333333333,
                                         '#9c179e'], [0.4444444444444444,
                                         '#bd3786'], [0.5555555555555556,
                                         '#d8576b'], [0.6666666666666666,
                                         '#ed7953'], [0.7777777777777778,
                                         '#fb9f3a'], [0.8888888888888888,
                                         '#fdca26'], [1.0, '#f0f921']],
                          'type': 'surface'}],
             'table': [{'cells': {'fill': {'color': '#EBF0F8'}, 'line': {'color': 'white'}},
                        'header': {'fill': {'color': '#C8D4E3'}, 'line': {'color': 'white'}},
                        'type': 'table'}]},
    'layout': {'annotationdefaults': {'arrowcolor': '#2a3f5f', 'arrowhead': 0, 'arrowwidth': 1},
               'autotypenumbers': 'strict',
               'coloraxis': {'colorbar': {'outlinewidth': 0, 'ticks': ''}},
               'colorscale': {'diverging': [[0, '#8e0152'], [0.1, '#c51b7d'],
                                            [0.2, '#de77ae'], [0.3, '#f1b6da'],
                                            [0.4, '#fde0ef'], [0.5, '#f7f7f7'],
                                            [0.6, '#e6f5d0'], [0.7, '#b8e186'],
                                            [0.8, '#7fbc41'], [0.9, '#4d9221'], [1,
                                            '#276419']],
                              'sequential': [[0.0, '#0d0887'],
                                             [0.1111111111111111, '#46039f'],
                                             [0.2222222222222222, '#7201a8'],
                                             [0.3333333333333333, '#9c179e'],
                                             [0.4444444444444444, '#bd3786'],
                                             [0.5555555555555556, '#d8576b'],
                                             [0.6666666666666666, '#ed7953'],
                                             [0.7777777777777778, '#fb9f3a'],
                                             [0.8888888888888888, '#fdca26'], [1.0,
                                             '#f0f921']],
                              'sequentialminus': [[0.0, '#0d0887'],
                                                  [0.1111111111111111, '#46039f'],
                                                  [0.2222222222222222, '#7201a8'],
                                                  [0.3333333333333333, '#9c179e'],
                                                  [0.4444444444444444, '#bd3786'],
                                                  [0.5555555555555556, '#d8576b'],
                                                  [0.6666666666666666, '#ed7953'],
                                                  [0.7777777777777778, '#fb9f3a'],
                                                  [0.8888888888888888, '#fdca26'],
                                                  [1.0, '#f0f921']]},
               'colorway': ["#db2b39","#3d405b","#2fbf71","#faa613","#00a6fb"],
               'font': {'color': '#2a3f5f'},
               'geo': {'bgcolor': 'white',
                       'lakecolor': 'white',
                       'landcolor': '#E5ECF6',
                       'showlakes': True,
                       'showland': True,
                       'subunitcolor': 'white'},
               'hoverlabel': {'align': 'left',
                             'namelength': 0},
               'hovermode': 'closest',
               'legend': {'orientation': 'v',
                          'bordercolor': '#3D3D3D',
                          'borderwidth': 0.7,
                          'itemwidth': 30,
                          'x': 0.01,
                          'y': 1.075,
                          'title': None,
                          'bgcolor':'rgba(255,255,255,0.70)'},
               'mapbox': {'style': 'light'},
               'paper_bgcolor': 'white',
               'plot_bgcolor': 'white',
               'polar': {'angularaxis': {'gridcolor': 'white', 'linecolor': 'white', 'ticks': ''},
                         'bgcolor': '#E5ECF6',
                         'radialaxis': {'gridcolor': 'white', 'linecolor': 'white', 'ticks': ''}},
               'scene': {'xaxis': {'backgroundcolor': '#E5ECF6',
                                   'gridcolor': 'white',
                                   'gridwidth': 2,
                                   'linecolor': 'white',
                                   'showbackground': True,
                                   'ticks': '',
                                   'zerolinecolor': 'white'},
                         'yaxis': {'backgroundcolor': '#E5ECF6',
                                   'gridcolor': 'white',
                                   'gridwidth': 2,
                                   'linecolor': 'white',
                                   'showbackground': True,
                                   'ticks': '',
                                   'zerolinecolor': 'white'},
                         'zaxis': {'backgroundcolor': '#E5ECF6',
                                   'gridcolor': 'white',
                                   'gridwidth': 2,
                                   'linecolor': 'white',
                                   'showbackground': True,
                                   'ticks': '',
                                   'zerolinecolor': 'white'}},
               'shapedefaults': {'line': {'color': '#2a3f5f'}},
               'ternary': {'aaxis': {'gridcolor': 'white', 'linecolor': 'white', 'ticks': ''},
                           'baxis': {'gridcolor': 'white', 'linecolor': 'white', 'ticks': ''},
                           'bgcolor': '#E5ECF6',
                           'caxis': {'gridcolor': 'white', 'linecolor': 'white', 'ticks': ''}},
               'title': {'x': 0.5,
                        'font_size':30},
               'xaxis': {'automargin': True,
                         'gridcolor': '#eeeeee',
                         'linecolor': 'white',
                         'tickformat': '%b\n%Y',
                         'ticks': '',
                         'title': {'text':None, 'standoff': 0, 'font': {'color':'white'}},
                         'zerolinecolor': 'white',
                         'zerolinewidth': 2,
                         'hoverformat': '%d %b %Y'},
               'yaxis': {'automargin': True,
                         'gridcolor': '#eeeeee',
                         'linecolor': 'white',
                         'ticks': '',
                         'title': {'text':None, 'standoff': 0, 'font': {'color':'white'}},
                         'zerolinecolor': 'white',
                         'zerolinewidth': 2}}
})


#Setting custom plotting theme as default

def set_theme():
    pio.templates['my_theme'] = my_theme
    pio.templates.default = 'my_theme'

set_theme()
config = {'displayModeBar':False}


fig = go.Figure()
fig.add_scatter(x=[6,6,8,8,6], y=[2,3,3,2,2], fill='toself', fillcolor='rgba(0,166,251,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[4,5,5,4,4], fill='toself', fillcolor='rgba(250,166,19,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[6,7,7,6,6], fill='toself', fillcolor='rgba(47,191,113,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[8,9,9,8,8], fill='toself', fillcolor='rgba(61,64,91,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[10,11,11,10,10], fill='toself', fillcolor='rgba(219,43,57,1)')
fig.add_scatter(x=[10],y=[5])
fig.add_scatter(x=[0],y=[0])
fig.add_annotation(x=3.5, y=2.5, text='Vaccinations    -', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=4.5, text='Tests              -', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=6.5, text='Recoveries      -', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=8.5, text='Deaths           -', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=10.5, text='Cases             -', font_size=35,showarrow=False)
fig.update_layout(xaxis_tickformat='', showlegend=False, xaxis_gridcolor='#ffffff', yaxis_gridcolor='#ffffff',
                 hoverlabel=None, margin=dict(t=90, b=0), xaxis_fixedrange=True, yaxis_fixedrange=True)
fig.update_xaxes(visible=False)
fig.update_yaxes(visible=False)
fig.update_traces(mode='text', hoverinfo='skip')
fig.add_annotation(x=0,y=1.15,showarrow=False,xref='paper', yref='paper', text='There are 5 main features in this \
project (listed below). Each feature is represented by its unique colour. And the',
                  font=dict(size=(15)))
fig.add_annotation(x=0,y=1.1, showarrow=False, xref='paper', yref='paper', text="plotting of the features will \
always follow the same order as below.<span style='font-size:0.65em'> (The legend order though is governed by ease of \
viewability)</span>",
                  font=dict(size=15))
fig.show(config=config)


#Function to get last valid value or last change in value

def last_valid_value(var, last_changed=False):
    val = var.iloc[var.last_valid_index()]
    count = 1
    while val == 0:
        val = var.iloc[var.last_valid_index() - count]
        count += 1
    if last_changed == True:
        while val == var.iloc[var.last_valid_index() - count]:
            count += 1
        else:
            return var.iloc[var.last_valid_index() - count]
    return val


fig = go.Figure()
fig.add_annotation(xref='paper',yref='paper',x=0.5,y=1.28, text='Today\'s Stats With Last Change', showarrow=False,
                  font=dict(size=30))
fig.add_indicator(mode = "number+delta", title='Total Cases',
    value = last_valid_value(world_df.total_cases),
    domain = {'x': [0, 0.25], 'y': [0.75, 1]},
    delta = {'reference': last_valid_value(world_df.total_cases,True), 'relative': False, 'position':'top'})
fig.add_indicator(mode = "number+delta", title='Total Deaths',
    value = last_valid_value(world_df.total_deaths),
    domain = {'x': [0.75, 1], 'y': [0.75, 1]},
    delta = {'reference': last_valid_value(world_df.total_deaths,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='Total Recoveries',
    value = last_valid_value(world_df.total_recovered),
    domain = {'x': [0.375, 0.625], 'y': [0.375, 0.625]},
    delta = {'reference': last_valid_value(world_df.total_recovered,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='Total Tests',
    value = last_valid_value(world_df.total_tests),
    domain = {'x': [0, 0.25], 'y': [0, 0.25]},
    delta = {'reference': last_valid_value(world_df.total_tests,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='Total Vaccinations',
    value = last_valid_value(world_df.total_vaccinations),
    domain = {'x': [0.75, 1], 'y': [0, 0.25]},
    delta = {'reference': last_valid_value(world_df.total_vaccinations,True), 'relative': False, 'position' : "top"})


fig = go.Figure()
fig.add_annotation(xref='paper',yref='paper',x=0.5,y=1.28, text='Today\'s Stats With Last Change', showarrow=False,
                  font=dict(size=30))
fig.add_indicator(mode = "number+delta", title='New Cases',
    value = last_valid_value(world_df.new_cases),
    domain = {'x': [0, 0.25], 'y': [0.75, 1]},
    delta = {'reference': last_valid_value(world_df.new_cases,True), 'relative': False, 'position':'top'})
fig.add_indicator(mode = "number+delta", title='New Deaths',
    value = last_valid_value(world_df.new_deaths),
    domain = {'x': [0.75, 1], 'y': [0.75, 1]},
    delta = {'reference': last_valid_value(world_df.new_deaths,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='New Recoveries',
    value = last_valid_value(world_df.new_recovered),
    domain = {'x': [0.375, 0.625], 'y': [0.375, 0.625]},
    delta = {'reference': last_valid_value(world_df.new_recovered,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='New Tests',
    value = last_valid_value(world_df.new_tests),
    domain = {'x': [0, 0.25], 'y': [0, 0.25]},
    delta = {'reference': last_valid_value(world_df.new_tests,True), 'relative': False, 'position' : "top"})
fig.add_indicator(mode = "number+delta", title='New Vaccinations',
    value = last_valid_value(world_df.new_vaccinations),
    domain = {'x': [0.75, 1], 'y': [0, 0.25]},
    delta = {'reference': last_valid_value(world_df.new_vaccinations,True), 'relative': False, 'position' : "top"})


fig = make_subplots(rows=1, cols=2, subplot_titles=['Normal Scale','Logarithmic Scale'])
[fig.add_scatter(x=world_df.date, y=world_df.total_cases, row=1, col=column+1, line=dict(color='#DB2B39'),
                 fill='tozeroy',hovertemplate='%{x}<br><b>%{y:.5s}</b><extra></extra>') for column in range(2)]
fig.update_yaxes(type="log", row=1, col=2)#, nticks=7)
fig.update_layout(title_text='Total Cases', showlegend=False)
fig.show(config=config)


fig = make_subplots(rows=1, cols=2, subplot_titles=['Normal Scale','Logarithmic Scale'])
[fig.add_scatter(x=world_df.date, y=world_df.total_deaths, row=1, col=column+1, line=dict(color='#3D405B'),
                fill='tozeroy', hovertemplate='%{x}<br><b>%{y:.4s}</b><extra></extra>') for column in range(2)]
fig.update_yaxes(type="log", row=1, col=2)
fig.update_layout(title_text='Total Deaths', showlegend=False)
fig.show(config=config)


world_df_rec = world_df.query('date < "08-05-2021"')

fig = make_subplots(rows=1, cols=2, subplot_titles=['Normal Scale','Logarithmic Scale'])
[fig.add_scatter(x=world_df_rec.date, y=world_df_rec.total_recovered, row=1, col=column+1, line=dict(color='#2FBF71'),
                 fill='tozeroy',hovertemplate='%{x}<br><b>%{y:.5s}</b><extra></extra>') for column in range(2)]
fig.update_yaxes(type="log", row=1, col=2)
fig.update_layout(title_text='Total Recoveries', showlegend=False)
fig.show(config=config)


fig = make_subplots(rows=1, cols=2, subplot_titles=['Normal Scale','Logarithmic Scale'])
[fig.add_scatter(x=world_df.date, y=world_df.total_tests, row=1, col=column+1, line=dict(color='#FAA613'),
                fill='tozeroy', hovertemplate='%{x}<br><b>%{y}</b><extra></extra>') for column in range(2)]
fig.update_yaxes(type="log", row=1, col=2)
fig.update_layout(title_text='Total Tests', showlegend=False)
fig.show(config=config)


world_df_vax = world_df.query("date > '2020-10-31'")

fig = make_subplots(rows=1, cols=2, subplot_titles=['Normal Scale','Logarithmic Scale'])
[fig.add_scatter(x=world_df_vax.date, y=world_df_vax.total_vaccinations, row=1, col=column+1,
                line=dict(color='#00A6FB'),fill='tozeroy', hovertemplate='%{x}<br><b>%{y}</b><extra></extra>')\
                                                                                 for column in range(2)]
fig.update_yaxes(type="log", row=1, col=2)
fig.update_layout(title_text='Total Vaccinations', showlegend=False)
fig.show(config=config)


#Function to pass custom legend names

def custom_legend_name(new_names, fig_in_func=None):
    if fig_in_func != None: #condition added for ML plotting
        for i, new_name in enumerate(new_names):
            fig_in_func.data[i].name = new_name
        return fig_in_func
    for i, new_name in enumerate(new_names):
        fig.data[i].name = new_name


fig = go.Figure()
[fig.add_scatter(x=world_df.date, y=world_df[c],) for c in ['new_cases','new_cases_smoothed']]
fig.update_layout(hoverlabel=dict(bgcolor='rgba(0,0,0,0.70)',font=dict(color='white')),title='New Cases',
                  hovermode='x unified',colorway=['#DB2B39','#0D0628'], height=640,
                  xaxis=dict(rangeslider=dict(visible=True),type="date"))

custom_legend_name(['New Cases','New Cases Smoothed'])
fig.show(config=config)


fig = go.Figure()
[fig.add_scatter(x=world_df.date, y=world_df[c],) for c in ['new_deaths','new_deaths_smoothed']]
fig.update_layout(hovermode='x unified', hoverlabel=dict(bgcolor='rgba(255,255,255,0.80)',font=dict(color='black')),
                 title='New Deaths',colorway=['#3D405B','#DC0073'],height=640,
                  xaxis=dict(rangeslider=dict(visible=True),type="date"))
custom_legend_name(['New Deaths','New Deaths Smoothed'])
fig.show(config=config)


fig = go.Figure()
[fig.add_scatter(x=world_df_rec.date, y=world_df_rec[c],) for c in ['new_recovered','new_recovered_smoothed']]
fig.update_layout(hovermode='x unified', hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')),
                 title='New Recoveries',colorway=['#2FBF71','#0D0628'],height=640,
                  xaxis=dict(rangeslider=dict(visible=True),type="date"))
custom_legend_name(['New Recovered','New Recovered Smoothed'])
fig.add_annotation(xref='paper',yref='paper',x=-0.05,y=-0.43,showarrow=False,
                  text="(The spike on 12th Dec is due to Turkey's govt providing a data dump on that very date. As \
the figures are official, I will make no changes to the graph.)")
fig.show(config=config)


world_df_tests = world_df[world_df['date'] < df.date.iloc[-2]]

fig = go.Figure()
[fig.add_scatter(x=world_df_tests.date, y=world_df_tests[c],) for c in ['new_tests','new_tests_smoothed']]
fig.update_layout(hovermode='x unified', hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')),
                 title='New Tests',colorway=['#FAA613','#28502E'],height=640,
                  xaxis=dict(rangeslider=dict(visible=True),type="date"))
custom_legend_name(['New Tests','New Tests Smoothed'])
fig.show(config=config)


fig = go.Figure()
[fig.add_scatter(x=world_df_vax.date, y=world_df_vax[c],) for c in ['new_vaccinations','new_vaccinations_smoothed']]
fig.update_layout(hovermode='x unified', hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',font=dict(color='white')),
                 title='New Vaccinations',colorway=['#00A6FB','#0D0628'],height=640,
                  xaxis=dict(rangeslider=dict(visible=True),type="date"))
custom_legend_name(['New Vaccinations','New Vaccinations Smoothed'])
fig.show(config=config)


fig = go.Figure()
fig.add_scatter(x=world_df.date, y=world_df.population, name='pop', line_color='#740667')
fig.add_scatter(x=world_df.date, y=world_df.total_cases, fill='tozeroy', mode='lines', line_color='#DB2B39')
fig.add_scatter(x=world_df.date, y=world_df.total_vaccinations, fill='tonexty', name='total_vax', line_color='#00A6FB')
fig.add_scatter(x=world_df.date, y=world_df.herd_immunity_threshold, name='HiT', line_color='#F65BE3')
fig.add_annotation(x=0.5, xref='paper',y = world_df.herd_immunity_threshold[0], text = 'Herd Immunity Threshold',
                  arrowhead=2, font=dict(color='#F65BE3'), arrowcolor='#F65BE3', arrowsize=1.2)
fig.add_annotation(x=0.25, xref='paper', y=world_df.population[0], text='Population', arrowhead=2,
                  font=dict(color='#740667'), arrowsize=1.2)
fig.add_annotation(x=0.15, xref='paper', y=round(len(world_df) * 0.15), text = 'Total Cases',
                  arrowhead=2, font=dict(color='#DB2B39'), arrowsize=1.2)
fig.add_annotation(x=0.92, xref='paper', y=world_df.total_vaccinations[round(len(world_df) * 0.92)],
                   text='Total Vaccinations', arrowhead=2, font=dict(color='#00A6FB'), arrowsize=1.2)
fig.update_layout(title_text='Herd Immunity Threshold', showlegend=False)
fig.update_traces(hovertemplate='%{x}<br><b>%{y}</b><extra></extra>')
fig.show(config=config)


fig = px.line(world_df_tests, x='date', y=['total_cases','total_deaths','total_recovered','total_tests',
                              'total_vaccinations'], log_y=True)
fig.update_traces(hovertemplate='%{y}')
fig.update_layout(legend=dict(title=None), title='Totals Comparison Line Chart', hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.5)',font=dict(color='black')))
custom_legend_name(['Total Cases','Total Deaths','Total Recovered','Total Tests','Total Vaccinations'])
fig.add_annotation(xref='paper',yref='paper',x=0,y=-0.16,showarrow=False,
                  text="(Testing had started in other countries before China started releasing official figures for\
 Cases. Hence, in the graph, Tests start before Cases)")
fig.show(config=config)


fig = px.area(world_df, x='date', y=['total_cases','total_deaths','total_recovered','total_tests',
                                     'total_vaccinations'])
fig.update_traces(hovertemplate='%{y}')
fig.update_layout(legend=dict(title=None, traceorder='reversed'), title='Totals Comparison Area Chart',
                  hovermode='x unified',hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',font=dict(color='white')))
custom_legend_name(['Total Cases','Total Deaths','Total Recovered','Total Tests','Total Vaccinations'])
fig.show(config=config)


fig = px.line(world_df, x='date', y=['new_cases','new_deaths','new_recovered','new_tests',
                                     'new_vaccinations'], log_y=True)
fig.update_traces(hovertemplate=None)
fig.update_layout(legend=dict(title=None), title='Daily Figures', hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')))
custom_legend_name(['New Cases','New Deaths','New Recovered','New Tests','New Vaccinations'])
fig.show(config=config)


fig = px.area(world_df_tests, x='date',y=['percent_pop_affected','percent_pop_tested','percent_pop_vaccinated'],
              title='Percent Population Area Chart', color_discrete_sequence=['#db2b39','#faa613','#00a6fb'])
fig.update_traces(hovertemplate=None, legendgroup=None)
fig.update_layout(legend=dict(title=None), hovermode='x unified',hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',
                                                                                  font=dict(color='white')))
custom_legend_name(['Percent Population Affected','Percent Population Tested','Percent Population Vaccinated'])
fig.show(config=config)


fig = px.line(world_df, x='date',y=['new_cases','hosp_patients','icu_patients'],title='Cases & Patients',
             color_discrete_sequence=['#db2b39','#390099','#00A676'])
fig.update_traces(hovertemplate=None, legendgroup=None)
fig.update_layout(legend=dict(title=None), hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')))
custom_legend_name(['New Cases','Hospital Patients','ICU Patients'])
fig.show(config=config)


fig = go.Figure()
fig.add_scatter(x=world_df.date, y=world_df.death_percent, fill='tozeroy', name='Case Fatality Percent',
               line=dict(color='#3d405b'))
fig.add_scatter(x=world_df.date, y=world_df.recovered_percent, fill='tonexty', name='Recovered Percent',
               line=dict(color='#2fbf71'))
fig.update_traces(hovertemplate=None, legendgroup=None)
fig.update_layout(legend=dict(title=None), hovermode='x unified', title='Recovered Percent & Case Fatality Percent',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')))
custom_legend_name(['Case Fatality Percent','Recovered Percent'])
fig.show(config=config)


fig= go.Figure()
fig.add_scatter(x=world_df.date, y=world_df.positive_rate, name='Positive Rate')
fig.add_scatter(x=world_df.date, y=world_df.tests_per_case, name='Tests Per Case', line=dict(color='#faa613'))
fig.update_layout(hovermode='x unified', hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',font=dict(color='white')),
                 title='Positive Rate & Tests Per Case')
fig.show(config=config)


fig = px.line(world_df, x='date', y='percent_increase_cases', color_discrete_sequence=['#F71735'],
        title='Growth Rate<br><sup><sup>(Percent change per day)</sup></sup>')
fig.update_traces(hovertemplate='%{x}<br>Percent change: %{y}')
fig.show(config=config)


#Calculating plot ratio of vaccination start day to total days

total_days = max(world_df.date) - min(world_df.date)
vax_start_day = pd.Timestamp('2020-12-13') - min(world_df.date)
vax_start_ratio = vax_start_day/total_days


fig = px.line(world_df, x='date', y='reproduction_rate', title='Coronavirus Reproduction Rate (R<sub>0</sub>)',
             color_discrete_sequence=['#F71735'])
fig.update_traces(hovertemplate='%{x}<br>R<sub>0</sub>: %{y}')
fig.add_shape(type='line', xref='paper',x0=vax_start_ratio ,x1=vax_start_ratio, y1=3.7)
fig.add_annotation(x=vax_start_ratio+0.01, xref='paper', y=0.75, yref='paper',axref='x domain',ayref='y domain',
                   ax=60, ay=-70,
                   text='Start of Vaccination<br>(13th Dec 2020)', arrowhead=2, arrowsize=1.2)
fig.show(config=config)


#Importing data for a graph
epidemic_df = pd.read_csv('https://raw.githubusercontent.com/vyaduvanshi/helper-files/master/epidemic_stats.csv')
epidemic_df.dropna(inplace=True)
epidemic_df.drop([2,3,7,12,13,15,17],inplace=True)

#Creating array for plotting line
x_array = []
for x,y in zip(epidemic_df.rnaught_lower,epidemic_df.rnaught_upper):
    x_array.append(x)
    x_array.append(y)   
y_array = []
for x,y in zip(epidemic_df.hit_lower,epidemic_df.hit_upper):
    y_array.append(x)
    y_array.append(y)

#Plotting graph
a = ['#23001E','#20A39E','#D00000','#29282A','#FF0A0A','#C6F91F','#FF3333','#AAF683','#FFBA49','#058C42']
fig = go.Figure()
for r0_low, r0_high, hit_low, hit_high in zip(epidemic_df.rnaught_lower, epidemic_df.rnaught_upper,
                                           epidemic_df.hit_lower, epidemic_df.hit_upper):
    fig.add_scatter(x=[r0_low,r0_low,0,0,r0_high,r0_high],y=[0,hit_low,hit_low,hit_high,hit_high,0], fill='toself',
                   mode='text')
fig.add_scatter(x=sorted(x_array), y=sorted(y_array), line_color='black', showlegend=False)
fig.update_xaxes(tickformat='', hoverformat='', title=dict(text='R<sub>0</sub>', standoff=15,
                                                           font=dict(color='#2A3F5F')))
fig.update_yaxes(title=dict(text='Herd Immunity Threshold', standoff=15, font=dict(color='#2A3F5F')))
fig.update_layout(showlegend=True, height=800, margin=dict(l=0,r=0), colorway = a,
                  title_text='R<sub>0</sub> vs Herd Immunity Threshold', legend_orientation='h', legend_borderwidth=0)
fig.update_traces(hovertemplate='<b>R<sub>0</sub>: </b>%{x}<br><b>HIT: </b>%{y}<extra></extra>')
custom_legend_name(list(epidemic_df.Disease), fig_in_func=fig)
fig.show()


fig = px.line(world_df, x='total_vaccinations_per_hundred', y='reproduction_rate', range_y=[0,1.4],
       title='Vaccination vs Reproduction (R<sub>0</sub>) Rates', color_discrete_sequence=['#F71735'])
fig.update_traces(hovertemplate='Total Vaccinations Per Hundred: %{x}<br>R<sub>0</sub>: %{y}')
fig.update_xaxes(title=dict(text='Total Vaccinations Per Hundred', standoff=15, font_color='#2a3f5f'),
                 hoverformat = '', tickformat = '')
fig.update_yaxes(title=dict(text='Reproduction Rate', standoff= 15, font_color='#2a3f5f'))
fig.show(config=config)


#Function to calculate rate of feature

def calc_rate(var, x, population):
    return round((var * x)/population,2)

pop = world_df.population[0]


#Data for table

header_vals = ['Feature','Per Capita','Per 100','Per 1000','Per Million']

values_world = [['Total Cases','Total Deaths','Total Recovered','Total Tests','Total Vaccinations',
                'People Vaccinated','People Fully Vaccinated'],
                
                [calc_rate(last_valid_value(world_df.total_cases),1,pop),
                 round(last_valid_value(world_df.total_deaths)/pop,4),
                 calc_rate(last_valid_value(world_df.total_recovered),1,pop),
                 calc_rate(last_valid_value(world_df.total_tests),1,pop),
                 calc_rate(last_valid_value(world_df.total_vaccinations),1,pop),
                 calc_rate(last_valid_value(world_df.people_vaccinated),1,pop),
                 calc_rate(last_valid_value(world_df.people_fully_vaccinated),1,pop)],
                
                [round(last_valid_value(world_df.total_cases_per_hundred),2),
                 round(last_valid_value(world_df.total_deaths_per_hundred),2),
                 round(last_valid_value(world_df.total_recovered_per_hundred),2),
                 round(last_valid_value(world_df.total_tests_per_hundred),2),
                 last_valid_value(world_df.total_vaccinations_per_hundred),
                 last_valid_value(world_df.people_vaccinated_per_hundred),
                 last_valid_value(world_df.people_fully_vaccinated_per_hundred)],
                
                [calc_rate(last_valid_value(world_df.total_cases),1000,pop),
                 calc_rate(last_valid_value(world_df.total_deaths),1000,pop),
                 calc_rate(last_valid_value(world_df.total_recovered),1000,pop),
                 calc_rate(last_valid_value(world_df.total_tests),1000,pop),
                 calc_rate(last_valid_value(world_df.total_vaccinations),1000,pop),
                 calc_rate(last_valid_value(world_df.people_vaccinated),1000,pop),
                 calc_rate(last_valid_value(world_df.people_fully_vaccinated),1000,pop)],
                
                [last_valid_value(world_df.total_cases_per_million),
                 last_valid_value(world_df.total_deaths_per_million),
                 calc_rate(last_valid_value(world_df.total_recovered),1000000,pop),
                 calc_rate(last_valid_value(world_df.total_tests),1000000,pop),
                 calc_rate(last_valid_value(world_df.total_vaccinations),1000000,pop),
                 calc_rate(last_valid_value(world_df.people_vaccinated),1000000,pop),
                 calc_rate(last_valid_value(world_df.people_fully_vaccinated),1000000,pop)]]


fig = go.Figure(data=[go.Table(
    header=dict(values=header_vals,
                fill_color='paleturquoise',
                align='center',
               height=35,
               font=dict(size=15)),
    cells=dict(values=values_world,
               fill_color='lavender',
               align='left',
              height=30,
              font=dict(size=13)))
])

fig.update_layout(title_text='Features in Various Rates',margin=dict(b=0))
fig.show(config=config)


cont_total_df = pd.pivot_table(continent_df, index='continent', aggfunc='last').reset_index()

for log_y in False,True:
    fig = go.Figure()
    fig = px.histogram(cont_total_df, x='continent', y=['total_cases','total_deaths','total_recovered','total_tests',
                                                        'total_vaccinations'], barmode='group', log_y=log_y,
                      title = '<sup>Logarithmic Scale<sup>' if log_y else 
                       'Continent Wise Comparison<br><sup>Normal Scale</sup>')
    fig.update_layout(legend=dict(orientation='h', title=None, borderwidth=0),title_font_size=30, title_y=0.94)
    fig.update_traces(hovertemplate='%{y}<extra></extra>')
    fig.update_xaxes(tickfont=dict(size=15))
    custom_legend_name(['Total Cases','Total Deaths','Total Recovered','Total Tests','Total Vaccinations'])
    fig.show(config=config)


fig = go.Figure()
fig.add_scatter(x=[6,6,8,8,6], y=[2,3,3,2,2], fill='toself', fillcolor='rgba(255,149,10,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[4,5,5,4,4], fill='toself', fillcolor='rgba(41,191,18,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[6,7,7,6,6], fill='toself', fillcolor='rgba(57,0,153,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[8,9,9,8,8], fill='toself', fillcolor='rgba(52,228,234,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[10,11,11,10,10], fill='toself', fillcolor='rgba(255,44,85,1)')
fig.add_scatter(x=[6,6,8,8,6], y=[12,13,13,12,12], fill='toself', fillcolor='rgba(83,134,228,1)')
fig.add_scatter(x=[10],y=[5])
fig.add_scatter(x=[0],y=[0])
fig.add_annotation(x=3.5, y=2.5, text='South America   -    ', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=4.5, text='Oceania             -    ', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=6.5, text='North America   -    ', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=8.5, text='Europe             -    ', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=10.5, text='Asia                 -    ', font_size=35,showarrow=False)
fig.add_annotation(x=3.5, y=12.5, text='Africa               -    ', font_size=35,showarrow=False)
fig.update_layout(xaxis_tickformat='', showlegend=False, xaxis_gridcolor='#ffffff', yaxis_gridcolor='#ffffff',
                 hoverlabel=None, margin=dict(t=90, b=0), xaxis_fixedrange=True, yaxis_fixedrange=True)
fig.update_xaxes(visible=False)
fig.update_yaxes(visible=False)
fig.update_traces(mode='text', hoverinfo='skip')
fig.add_annotation(x=0,y=1.15,showarrow=False,xref='paper', yref='paper', text='The continents are represented by \
the following colours. Their order of plotting will always be alphabetical.',
                  font=dict(size=(15)))
fig.show(config=config)


#Passing colours for each continent

pio.templates['my_theme'].layout.colorway= ["#5386E4","#ff2c55","#34e4ea","#390099","#29bf12","#ff950a"]


fig = px.area(continent_df, x='date', y='new_cases_smoothed', color='continent',
             title='New Cases', groupnorm='percent', hover_data=['new_cases'], height=550)
fig.update_traces(hovertemplate='%{customdata:,}', legendgroup=None, mode='none')
fig.update_layout(hovermode='x unified', legend=dict(title=None, orientation='h', borderwidth=0),
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.4)',font=dict(color='white')))
fig.show(config=config)


fig = px.area(continent_df, x='date', y='new_deaths_smoothed', color='continent',
             title='New Deaths', groupnorm='percent', hover_data=['new_deaths'], height=550)
fig.update_traces(hovertemplate='%{customdata:,}', mode='none', legendgroup=None)
fig.update_layout(hovermode='x unified', legend=dict(title=None, orientation='h', borderwidth=0),
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.4)',font=dict(color='white')))
fig.show(config=config)


continent_df_rec = continent_df.query('date < "08-05-2021"')

fig = px.area(continent_df_rec, x='date', y='new_recovered_smoothed', color='continent',
             title='New Recoveries', groupnorm='percent', hover_data=['new_recovered'], height=550)
fig.update_traces(hovertemplate='%{customdata:,}', mode='none', legendgroup=None)
fig.update_layout(hovermode='x unified', legend=dict(title=None, orientation='h', borderwidth=0),
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.4)',font=dict(color='white')))
fig.show(config=config)


fig = px.area(continent_df, x='date', y='new_tests_smoothed', color='continent',
             title='New Tests', groupnorm='percent', hover_data=['new_tests'], height=550)
fig.update_traces(hovertemplate='%{customdata:,}', mode='none', legendgroup=None)
fig.update_layout(hovermode='x unified', legend=dict(title=None, orientation='h', borderwidth=0),
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.4)',font=dict(color='white')))
fig.show(config=config)


continent_df_vax = continent_df.query("date>'2020-11-20'")
continent_df_vax = continent_df_vax[continent_df_vax['date'] < df.date.iloc[-1]]

fig = px.area(continent_df_vax, x='date', y='new_vaccinations_smoothed', color='continent',
             title='New Vaccinations', groupnorm='percent', hover_data=['new_vaccinations'], height=550)
fig.update_traces(hovertemplate='%{customdata:,}', mode='none', legendgroup=None)
fig.update_layout(hovermode='x unified', legend=dict(title=None, orientation='h', borderwidth=0),
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.4)',font=dict(color='white')))
fig.show(config=config)


domains = [
    {'x': [0.0, 0.38], 'y': [0.0, 0.38]}, #bot left
    {'x': [0.31, 0.69], 'y': [0.31, 0.69]}, # mid
    {'x': [0.0, 0.38], 'y': [0.62, 1.0]}, # top left
    {'x': [0.62, 1], 'y': [0.0, 0.38]}, #bot right
    {'x': [0.62, 1], 'y': [0.62, 1.00]}, #top right
]

fig = go.Figure()
fig.add_pie(labels=continent_df.continent, values=continent_df.new_tests, domain=domains[0], title_text='Tests')
fig.add_pie(labels=continent_df.continent, values=continent_df.new_recovered, domain=domains[1],
            title_text='Recoveries')
fig.add_pie(labels=continent_df.continent, values=continent_df.new_cases, domain=domains[2], title_text='Cases')
fig.add_pie(labels=continent_df.continent, values=continent_df.new_vaccinations, domain=domains[3],
            title_text='Vaccinations')
fig.add_pie(labels=continent_df.continent, values=continent_df.new_deaths, domain=domains[4], title_text='Deaths')

fig.update_layout(width=950, height=800, legend={'x':-0.09,'y':0.52}, title='Continent Comparison Across Features \
(Totals)')
fig.update_traces(hoverinfo="label+value", textinfo='percent', opacity=0.8, sort=False, direction='clockwise',
                  insidetextorientation='horizontal', title_position='bottom center', title_font_size=21,
                  marker=dict(line=dict(color='#000000',width=0.6)))
fig.show(config=config)


temp_df = continent_df.copy()

for var in ['new_cases_smoothed','new_deaths_smoothed','new_recovered_smoothed','new_tests_smoothed',
         'new_vaccinations_smoothed']:
    if var == 'new_vaccinations_smoothed':
        temp_df = continent_df_vax
    if var == 'new_recovered_smoothed':
        temp_df = continent_df_rec
    fig = px.line(temp_df, x='date', y=var, color='continent', title=var.replace('_',' ').title().
                  replace('Smoothed','Line Chart').replace('New','Daily'))
    fig.update_traces(hovertemplate=None, legendgroup=None)
    fig.update_layout(legend=dict(title=None,orientation='v'), hovermode='x unified',
                     hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',font=dict(color='white')))
    fig.show(config=config)


for continent in continent_list:
    temp_df = continent_df_vax[continent_df_vax['continent']==continent]
    fig = go.Figure()
    fig.add_scatter(x=continent_df_vax.date,
                    y=temp_df.people_fully_vaccinated,fill='tozeroy', fillcolor='rgba(0,166,251,0.8)',
                    line_color='rgba(0,166,251,0.8)', name='Fully vaccinated',
                    hovertemplate='%{y:,.4s}<br>% of population : '+'%{text:.2f}',
                   text=(temp_df.people_fully_vaccinated/temp_df.population *100))
    fig.add_scatter(x=continent_df_vax.date,
                    y=temp_df.people_vaccinated,fill='tonexty', fillcolor='rgba(0,166,251,0.3)',
                    line_color='rgba(0,166,251,0.9)',name='At least 1 dose',
                   hovertemplate='%{y:,.4s}<br>% of population : '+'%{text:.2f}',
                   text=(temp_df.people_vaccinated/temp_df.population *100))
    fig.update_layout(title_text='Vaccinations<br><sub>'+continent+'</sub>', legend=dict(orientation='h',
                                                                                        borderwidth=0),
                     hovermode='x unified', hoverlabel=dict(bgcolor='rgba(255,255,255,0.5)',font=dict(color='black')))
    if continent == 'Asia':
        fig.add_annotation(xref='paper',yref='paper',x=0,y=-0.2, showarrow=False,
                           text="<span style='font-size:1em;color:dark-gray'>(The spike on 10th June is \
due to China\'s sudden reporting on that date. As the numbers are official, \
I will make no changes to the graph.)</span>")
    if continent == 'Oceania':
        fig.add_annotation(xref='paper',yref='paper',x=-0.1,y=-0.2, showarrow=False,
                           text="<span style='font-size:1em;color:dark-gray'>(The spike on 24th May is \
due to Australia\'s commencement of reporting for second doses. As the numbers are official, \
I will make no changes to the graph.)</span>")
    fig.show(config=config)


fig = px.area(continent_df, x='date', y='death_percent', color='continent',title='Case Fatality Percent')
fig.update_layout(legend=dict(orientation='h',borderwidth=0, title=None), hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.7)',font=dict(color='black')))
fig.update_traces(mode='none',hovertemplate='%{y:.2f}', legendgroup=None)
fig.show(config=config)


fig = px.area(continent_df_rec, x='date', y='recovered_percent', color='continent',title='Recovered Percent')
fig.update_layout(legend=dict(orientation='h',borderwidth=0, title=None), hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(255,255,255,0.7)',font=dict(color='black')))
fig.update_traces(mode='none',hovertemplate='%{y:.2f}', legendgroup=None)
fig.show(config=config)


for feature in ('percent_pop_affected','percent_pop_tested','percent_pop_vaccinated'):
    fig = px.area(continent_df_vax if feature == 'percent_pop_vaccinated' else continent_df,
                  x='date', y=feature, color='continent',
                  title=feature.replace('pop','Population').replace('_',' ').title())
    fig.update_layout(legend=dict(orientation='h',borderwidth=0, title=None), hovermode='x unified',
                     hoverlabel=dict(bgcolor='rgba(255,255,255,0.7)',font=dict(color='black')))
    fig.update_traces(mode='none',hovertemplate='%{y:.2f}', legendgroup=None)
    fig.show(config=config)


fig = px.line(continent_df, x='date', y='tests_per_case_smoothed', color='continent',title='Tests Per Case',
             log_y=True)
fig.update_layout(legend_title=None, hovermode='x unified',
                 hoverlabel=dict(bgcolor='rgba(0,0,0,0.75)',font=dict(color='white')))
fig.update_traces(hovertemplate='%{y:.2f}', legendgroup=None)
fig.show(config=config)


def max_range_aniframe(feature):
    return continent_df_25d[feature].max() + (continent_df_25d[feature].max()/10)


fig = px.bar(continent_df_25d, x='continent', y='total_cases',animation_frame=continent_df_25d.date.astype(str),
             log_y=False, range_y=[1,max_range_aniframe('total_cases')], color_discrete_sequence=['#db2b39'],
             title='Total Cases', text='total_cases')

annotation_date = list(continent_df_25d.date.sort_values().dt.strftime('%d %b %Y'))
fig.update_layout(showlegend=False, xaxis_tickfont_size=15)
fig.update_traces(customdata= annotation_date[0:6], texttemplate='%{y:,}', textposition='outside',
                  hovertemplate='%{customdata}<br>%{y}')
fig.add_annotation(x=0.5,y=1.15,text=annotation_date[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(customdata=annotation_date[0+(6*idx):6+(6*idx)],
                         hovertemplate='%{customdata}<br>%{y}', textposition='outside')
    frame.layout.update(annotations=[dict(x=0.5,y=1.15,text=annotation_date[idx*6],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


fig = px.bar(continent_df_25d, x='continent', y='total_deaths',animation_frame=continent_df_25d.date.astype(str),
             log_y=False, range_y=[1,max_range_aniframe('total_deaths')], color_discrete_sequence=['#3d405b'],
             title='Total Deaths', text='date')

fig.update_layout(showlegend=False, xaxis_tickfont_size=15)
fig.update_traces(customdata= annotation_date[0:6], texttemplate='%{y:,}', textposition='outside',
                  hovertemplate='%{customdata}<br>%{y}')
fig.add_annotation(x=0.5,y=1.15,text=annotation_date[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(customdata=annotation_date[0+(6*idx):6+(6*idx)],
                         hovertemplate='%{customdata}<br>%{y}', textposition='outside')
    frame.layout.update(annotations=[dict(x=0.5,y=1.15,text=annotation_date[idx*6],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


continent_df_25d_rec = continent_df_25d.query('date < "08-05-2021"')

fig = px.bar(continent_df_25d_rec, x='continent', y='total_recovered',
             animation_frame=continent_df_25d_rec.date.astype(str), log_y=False,
             range_y=[1,max_range_aniframe('total_recovered')], color_discrete_sequence=['#2fbf71'],
             title='Total Recovered', text='date')

annotation_date_rec = list(continent_df_25d_rec.date.sort_values().dt.strftime('%d %b %Y'))
fig.update_layout(showlegend=False, xaxis_tickfont_size=15)
fig.update_traces(customdata= annotation_date_rec[0:6], texttemplate='%{y:,}', textposition='outside',
                  hovertemplate='%{customdata}<br>%{y}')
fig.add_annotation(x=0.5,y=1.15,text=annotation_date_rec[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(customdata=annotation_date_rec[0+(6*idx):6+(6*idx)],
                         hovertemplate='%{customdata}<br>%{y}', textposition='outside')
    frame.layout.update(annotations=[dict(x=0.5,y=1.15,text=annotation_date_rec[idx*6],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


fig = px.bar(continent_df_25d, x='continent', y='total_tests',animation_frame=continent_df_25d.date.astype(str),
             log_y=False, range_y=[1,max_range_aniframe('total_tests')], color_discrete_sequence=['#faa613'],
             title='Total Tests', text='date')

fig.update_layout(showlegend=False, xaxis_tickfont_size=15)
fig.update_traces(customdata= annotation_date[0:6], texttemplate='%{y:,}', textposition='outside',
                  hovertemplate='%{customdata}<br>%{y}')
fig.add_annotation(x=0.5,y=1.15,text=annotation_date[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(customdata=annotation_date[0+(6*idx):6+(6*idx)],
                         hovertemplate='%{customdata}<br>%{y}', textposition='outside')
    frame.layout.update(annotations=[dict(x=0.5,y=1.15,text=annotation_date[idx*6],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


continent_df_25d_vax = continent_df_25d.query('date > "27-10-2020"')

fig = px.bar(continent_df_25d_vax, x='continent', y='total_vaccinations',
             animation_frame=continent_df_25d_vax.date.astype(str), log_y=False,
             range_y=[1,max_range_aniframe('total_vaccinations')], color_discrete_sequence=['#00a6fb'],
             title='Total Vaccinations', hover_name='date', text='date')

annotation_date_vax = list(continent_df_25d_vax.date.sort_values().dt.strftime('%d %b %Y'))
fig.update_layout(showlegend=False, xaxis_tickfont_size=15)
fig.update_traces(customdata= annotation_date_vax[0:6], texttemplate='%{y:,}', textposition='outside',
                  hovertemplate='%{customdata}<br>%{y}')
fig.add_annotation(x=0.5,y=1.15,text=annotation_date_vax[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(customdata=annotation_date_vax[0+(6*idx):6+(6*idx)],
                         hovertemplate='%{customdata}<br>%{y}', textposition='outside')
    frame.layout.update(annotations=[dict(x=0.5,y=1.15,text=annotation_date_vax[idx*6],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


#Function to change SI-prefix units from machine to human

def human_format(num):
    num = float('{:.3g}'.format(num))
    magnitude = 0
    while abs(num) >= 1000:
        magnitude += 1
        num /= 1000.0
    return '{}{}'.format('{:f}'.format(num).rstrip('0').rstrip('.'), 
                         ['', 'K', 'M', 'B', 'T'][magnitude])


#Creating two variables for 3d plots

colorlist = list(np.arange(0,continent_df_25d.shape[0]/6)) * 6
datelist = [str(x) if idx%4==0 else '' for idx,x in enumerate(continent_df_25d.date.dt.strftime('%d %b %Y').unique())]


continent_df_25d["hover_data"] = continent_df_25d.apply(lambda r: 
    f"Tests: {human_format(r['total_tests'])}<br>Vaccinations: {human_format(r['total_vaccinations'])}",axis=1)

annotation_dates_3d = list(continent_df_25d.date.dt.strftime('%d %b %Y'))

fig = px.scatter_3d(continent_df_25d, x='total_cases', y='total_deaths', z='total_tests',
             color=colorlist, symbol='continent', text=annotation_dates_3d,
                    symbol_sequence=['square','diamond','circle','x','cross','square-open'],
                   custom_data=['hover_data'], title='Cases, Deaths, Tests', opacity=0.85,
                   color_continuous_scale='Turbo', labels={'total_tests':'Total Tests',
                                                          'total_deaths':'Total Deaths',
                                                          'total_cases':'Total Cases'})

camera = dict(eye=dict(x=1.75, y=-1.25, z=0.65))
                                                        
fig.update_layout(margin=dict(l=0, r=0, b=0, t=100), scene_camera=camera, legend=dict(orientation='h',borderwidth=0,
                                                                                   title_text=None),
                 coloraxis_colorbar=dict(title=None, thickness=20, len=1, tickvals=np.arange(0,len(datelist)),
                                    ticktext=datelist))
fig.update_traces( mode='markers', hovertemplate='<b>%{text}</b><br>Cases: %{x}<br>Deaths: %{y}<br>%{customdata}')
fig.show(config=config)


fig = px.scatter_3d(continent_df_25d, x='total_cases_per_hundred', y='total_deaths_per_hundred',
                    z='total_tests_per_hundred', color = colorlist, symbol='continent', text=annotation_dates_3d,
                   symbol_sequence=['square','diamond','circle','x','cross','square-open'],
                   custom_data=['total_vaccinations_per_hundred'], title='(Cases, Deaths, Tests) Per Hundred \
Population',
                    opacity=0.85, color_continuous_scale='Turbo', labels={'total_tests_per_hundred':'Total Tests',
                                                          'total_deaths_per_hundred':'Total Deaths',
                                                          'total_cases_per_hundred':'Total Cases'})

camera = dict(eye=dict(x=1.75, y=-1.25, z=0.65))

fig.update_layout(margin=dict(l=0, r=0, b=0, t=100), scene_camera=camera, legend=dict(orientation='h',borderwidth=0,
                                                                                   title_text=None),
                 coloraxis_colorbar=dict(title=None, thickness=20, len=1, tickvals=np.arange(0,len(datelist)),
                                    ticktext=datelist))
fig.update_traces( mode='markers', hovertemplate='<b>%{text}</b><br>Cases: %{x}<br>Deaths: %{y}<br>\
Tests: %{z}<br>Vaccinations: %{customdata}')
fig.show(config=config)


#Creating temporary dataframes for plotting of all five features

temp_df = pd.pivot_table(df, index=['continent','country'], aggfunc='last').reset_index()
temp_df['world'] = 'World'
temp_df_cases = temp_df[temp_df['total_cases'].notna()]
temp_df_deaths = temp_df[temp_df['total_deaths'].notna()]
temp_df_recovered = temp_df[temp_df['total_recovered'].notna()]
temp_df_recovered = temp_df_recovered[(temp_df_recovered['total_recovered'] != 0)]
temp_df_tests = temp_df[temp_df['total_tests'].notna()]
temp_df_vax = temp_df[temp_df['total_vaccinations'].notna()]
# temp_df_vax = temp_df_vax[(temp_df_vax['total_vaccinations'])]

temp_df_cases['total_cases_per_hundred'] = round(temp_df_cases['total_cases_per_hundred'],4)
temp_df_deaths['total_deaths_per_hundred'] = round(temp_df_deaths['total_deaths_per_hundred'],5)
temp_df_tests['total_tests_per_hundred'] = round(temp_df_tests['total_tests_per_hundred'],3)
temp_df_vax['total_vaccinations_per_hundred'] = round(temp_df_vax['total_vaccinations_per_hundred'],3)


fig = px.treemap(temp_df_cases, path=['world','continent','country'], values='total_cases', color='total_cases',
                 color_continuous_scale=["#FCEEEF", "#db2b39"], title='Total Cases', hover_name='country')

fig.update_traces(hovertemplate=None, hoverinfo='label+value+percent root+percent parent')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.show(config=config)


fig = px.treemap(temp_df_cases, path=['world','continent','country'], values='total_cases_per_hundred',
                 color='total_cases_per_hundred', title='Total Cases Per 100 Population',
                 color_continuous_scale=["#FCEEEF", "#db2b39"], custom_data=['total_cases_per_hundred'])

fig.update_traces(hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.data[0].customdata[-1] = round(last_valid_value(world_df['total_cases_per_hundred']),3)
fig.data[0].customdata[-7:-1] = np.round(cont_total_df[['total_cases_per_hundred']].values,3)
    
fig.show(config=config)


fig = px.treemap(temp_df_deaths, path=['world','continent','country'], values='total_deaths',
          color='total_deaths', color_continuous_scale=["#ffffff", "#3d405b"], title='Total Deaths')

fig.update_traces(hovertemplate=None, hoverinfo='label+value+percent root+percent parent')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.show(config=config)


fig = px.treemap(temp_df_deaths, path=['world','continent','country'], values='total_deaths_per_hundred',
                 color='total_deaths_per_hundred', title='Total Deaths Per 100 Population',
                 color_continuous_scale=["#ffffff", "#3d405b"], custom_data=['total_deaths_per_hundred'])

fig.update_traces(hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.data[0].customdata[-1] = round(last_valid_value(world_df['total_deaths_per_hundred']),3)
fig.data[0].customdata[-7:-1] = np.round(cont_total_df[['total_deaths_per_hundred']].values,3)
    
fig.show(config=config)


fig = px.treemap(temp_df_recovered, path=['world','continent','country'], values='total_recovered',
          color='total_recovered', color_continuous_scale=["#ffffff", "#2fbf71"], title='Total Recovered')

fig.update_traces(hovertemplate=None, hoverinfo='label+value+percent root+percent parent')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.show(config=config)


fig = px.treemap(temp_df_recovered, path=['world','continent','country'], values='total_recovered_per_hundred',
                 color='total_recovered_per_hundred', title='Total Recovered Per 100 Population',
                 color_continuous_scale=["#ffffff", "#2fbf71"], custom_data=['total_recovered_per_hundred'])

fig.update_traces(hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.update_layout( margin = dict(t=80, l=0, r=0, b=10), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.data[0].customdata[-1] = round(last_valid_value(world_df['total_recovered_per_hundred']),3)
fig.data[0].customdata[-7:-1] = np.round(cont_total_df[['total_recovered_per_hundred']].values,3)
    
fig.show(config=config)


fig = px.sunburst(temp_df_tests, path=['world','continent','country'], values='total_tests',
          color='total_tests', color_continuous_scale=["#ffffff", "#faa613"], title='Total Tests')

fig.update_traces(hovertemplate=None, hoverinfo='label+value+percent root+percent parent')
fig.update_layout( margin = dict(t=50, l=0, r=0, b=0), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.show(config=config)


fig = px.sunburst(temp_df_tests, path=['world','continent','country'], values='total_tests_per_hundred',
                 color='total_tests_per_hundred', title='Total Tests Per 100 Population',
                 color_continuous_scale=["#ffffff", "#faa613"], custom_data=['total_tests_per_hundred'])

fig.update_traces(hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.update_layout( margin = dict(t=50, l=0, r=0, b=0), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.data[0].customdata[-1] = round(last_valid_value(world_df['total_tests_per_hundred']),3)
fig.data[0].customdata[-7:-1] = np.round(cont_total_df[['total_tests_per_hundred']].values,3)
    
fig.show(config=config)


#Creating and filling values into variables to pass to vaccination sunburst plot

temp_list = [['People Vaccinated ' + country, 'People Fully Vaccinated ' + country] for country in temp_df_vax.country]
temp_list = [x for y in temp_list for x in y]
ids = ['World'] + list(temp_df_vax['continent'].unique()) + list(temp_df_vax['country']) + temp_list
country_end = next(x for x in ids if 'Vaccinated' in x)
country_end_index = ids.index(country_end)
labels = ids[0:country_end_index] + ['People Vaccinated', 'People Fully Vaccinated'] * len(ids[7:country_end_index])
parents = []
values = []
values_per_100 = []
values_percent_world = []
for x in ids:
    if x == 'World':
        parents.append('')
        values.append(0)
        values_per_100.append(last_valid_value(world_df.total_vaccinations_per_hundred))
    elif x in list(temp_df_vax.continent):
        parents.append('World')
        values.append(0)
        values_per_100.append(cont_total_df[cont_total_df.continent == x].total_vaccinations_per_hundred.values[0])
    elif x in list(temp_df_vax.country):
        parents.append(temp_df_vax[temp_df_vax.country == x].continent.values[0])
        values.append(temp_df_vax[temp_df_vax.country == x].total_vaccinations.values[0])
        values_per_100.append(temp_df_vax[temp_df_vax.country == x].total_vaccinations_per_hundred.values[0])
    elif 'Fully' in x:
        x = x[24:]
        parents.append(x)
        values.append(temp_df_vax[temp_df_vax.country == x].people_fully_vaccinated.values[0])
        values_per_100.append(temp_df_vax[temp_df_vax.country == x].people_fully_vaccinated_per_hundred.values[0])
    elif 'People Vaccinated' in x:
        x = x[18:]
        parents.append(x)
        values.append(temp_df_vax[temp_df_vax.country == x].people_vaccinated.values[0])
        values_per_100.append(temp_df_vax[temp_df_vax.country == x].people_vaccinated_per_hundred.values[0])


#Creating custom colorscale for vaccination sunburst graphs

a = n_colors('rgb(255, 255, 255)','rgb(0,166,251)', 256, colortype='rgb')
b = []
for x in a:
    b.append(x)
    b.append(x)
b.insert(0, 'rgb(128,210,253)')


#Creating a variable to show values in hoverlabel

values2 = values.copy()
values2[0] = last_valid_value(world_df.total_vaccinations)
values2[1:7] = cont_total_df['total_vaccinations'].values


#Calculating relative values from parent nodes to show in hoverlabels

percent_from_world = []
percent_from_parent = []

for x in range(len(ids)):
    if x < 217:
        percent_from_world.append(round(values2[x]/values2[0]*100,3))
    else:
        percent_from_world.append(np.nan)

for x in range(len(ids)):
    if parents[x] == 'Africa':
        percent_from_parent.append(round(values2[x]/values2[1]*100,3))
    elif parents[x] == 'Asia':
        percent_from_parent.append(round(values2[x]/values2[2]*100,3))
    elif parents[x] == 'Europe':
        percent_from_parent.append(round(values2[x]/values2[3]*100,3))
    elif parents[x] == 'North America':
        percent_from_parent.append(round(values2[x]/values2[4]*100,3))
    elif parents[x] == 'Oceania':
        percent_from_parent.append(round(values2[x]/values2[5]*100,3))
    elif parents[x] == 'South America':
        percent_from_parent.append(round(values2[x]/values2[6]*100,3))
    else:
        percent_from_parent.append(np.nan)


#Creating a column in dataframe where each row consists of a string to be displayed as hoverlabel

temp_df_values2 = pd.DataFrame(values2)
temp_df_values2['percent_from_world'] = percent_from_world
temp_df_values2['percent_from_parent'] = percent_from_parent
temp_df_values2['parent'] = parents
temp_df_values2["hover_data"] = temp_df_values2.apply(lambda r: 
    f"Vaccinations: {human_format(r[0])}<br>% of {r['parent']}: {r['percent_from_parent']}<br>\
% of World: {r['percent_from_world']}",axis=1)
temp_df_values2.hover_data.iloc[217:] = temp_df_values2.iloc[217:].hover_data.str.split('<br>').str[0]
temp_df_values2.hover_data.iloc[0] = temp_df_values2.hover_data.iloc[0].replace('% of : nan<br>','')
temp_df_values2.hover_data.iloc[1:7] = temp_df_values2.hover_data.iloc[1:7].str.replace('% of World: nan<br>','')


fig = px.sunburst(names=labels, parents=parents, values=values, ids=ids, color=values,
                 color_continuous_scale=b, title='Total Vaccinations')

fig.update_layout(margin = dict(t=50, l=0, r=0, b=0), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))
fig.update_traces(customdata=temp_df_values2['hover_data'],
                  hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.show(config=config)


fig = px.sunburst(names=labels, parents=parents, values=values_per_100, ids=ids, color=values_per_100,
                 color_continuous_scale=b, title='Total Vaccinations Per 100 Population')

fig.update_traces(customdata=temp_df_vax.total_vaccinations_per_hundred,
                  hovertemplate='<b>%{label}</b><br>%{customdata}')
fig.update_layout( margin = dict(t=50, l=0, r=0, b=0), coloraxis_colorbar=dict(title=None,thickness=20,
                                                                               len=1))    
fig.show(config=config)


#Modifying default plotting options for all further graphs

pio.templates['my_theme'].layout.xaxis.title = {'standoff': 15}
pio.templates['my_theme'].layout.yaxis.title = {'standoff': 15}
pio.templates['my_theme'].layout.plot_bgcolor = '#f8f9fa'
pio.templates['my_theme'].layout.xaxis.tickformat = ''
pio.templates['my_theme'].layout.xaxis.hoverformat = ''


#Creating a dataframe with mean values of each column for each country

df_avg_value = pd.pivot_table(df, index=['country','continent'], aggfunc='mean').reset_index()


#Function to plot two variables with their linear trendline

def plot_with_trendline(x, y, trendline=False, log_x=False, log_y=False):
    if trendline:
        fig = px.scatter(df_avg_value, x=x, y=y, trendline='ols', log_x=(True if log_x else False),
                        log_y=(True if log_y else False))
        fig.data[1]['line']['color'] = '#212529'
        trendline = fig.data[1]
        trendline_split = trendline.hovertemplate.split('<br>')
        trendline.hovertemplate = (trendline_split[0] +
#         '<br>' + trendline_split[1].replace('_',' ').title() +         #Comment out to add y = mx+c in hoverlabel
        '<br>' + trendline_split[2])
    
    fig = px.scatter(df_avg_value, x=x, y=y, color='continent', custom_data=['country'],
                     log_x=(True if log_x else False),log_y=(True if log_y else False),
                     category_orders={'continent':['Africa','Asia','Europe','North America','Oceania',
                                                   'South America']})
    x = x.replace('_',' ').title()
    y = y.replace('_',' ').title()
    fig.update_xaxes(title=x)
    fig.update_yaxes(title=y)
    fig.update_traces(hovertemplate='%{customdata}<br>'+x+': %{x}<br>'+y+': %{y}', marker_size=8)
    if trendline:
        fig.add_trace(trendline)
    fig.update_layout(legend=dict(orientation='h', title=None, borderwidth=0))
    return fig


#Function to plot more than two variables with linear trendlines for each relation

def plot_multiple_trendlines(x, y, xy_reverse=False):
    x_legend = []
    if xy_reverse:
        fig = px.scatter(df_avg_value, x=y, y=x, trendline='ols', custom_data=['continent','country'])
    else:
        fig = px.scatter(df_avg_value, x=x, y=y, trendline='ols', custom_data=['continent','country'])
    y = y.replace('_',' ').title()
    
    for var in range(len(x)):
        trendline = fig.data[var*2+1]
        trendline_split = trendline.hovertemplate.split('<br>')
        trendline.hovertemplate = (trendline_split[0] +
#         '<br>' + x[1].replace('_',' ').title() +            #Comment out to add y = mx+c in hoverlabel
        '<br>' + trendline_split[2])
        if x[var][:3] == 'gm_':
            x_element = x[var][3:].replace('_',' ').title()
        elif x[var][-5:] == 'Index':
            x_element = "".join([(" "+i if i.isupper() else i) for i in x[var]]).strip()
        else:
            x_element = x[var].replace('_',' ').title()
        
        x_legend = x_legend + [x_element] + ['']
        fig.data[var*2].hovertemplate = ('%{customdata[1]}<br>'+x_element+': %{x}<br>'+y+': %{y}\
<extra>%{customdata[0]}</extra>') if not xy_reverse else ('%{customdata[1]}<br>'+x_element+': %{y}<br>'+y+': %{x}\
<extra>%{customdata[0]}</extra>')
        
    fig.update_xaxes(title=y) if xy_reverse else fig.update_xaxes(title='Value')
    fig.update_yaxes(title='Value') if xy_reverse else fig.update_yaxes(title=y)  
    fig.update_traces(marker_size=8)
    fig.update_layout(legend=dict(orientation='h', title=None, borderwidth=0))
    for i, new_name in enumerate(x_legend):
        fig.data[i].name = new_name
    return fig


fig = plot_with_trendline('median_age', 'total_deaths_per_hundred', True)
fig.show(config=config)


fig = plot_multiple_trendlines(['aged_70_older','aged_65_older'], 'total_deaths_per_hundred')
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Percentages of population)')
fig.show(config=config)


fig = plot_with_trendline('diabetes_prevalence','total_deaths_per_hundred',True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(% of population aged 20 to 79 in 2017)')
fig.show(config=config)


fig = plot_with_trendline('cardiovasc_death_rate','total_deaths_per_hundred', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Annual deaths per 100k people, 2017)')
fig.show(config=config)


fig = plot_multiple_trendlines(['male_smokers','female_smokers'], 'total_deaths_per_hundred')
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Percentages of population)')
fig.show(config=config)


fig = plot_with_trendline('aged_65_older','Protection of elderly people',True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Percentage of population)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Protection of Elderly People- 0=No measures, 3=Extensive restrictions for isolation and\
 hygeine)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('hospital_beds_per_thousand','death_percent', True)
fig.update_yaxes(title='Case Fatality Percent')
fig.show(config=config)


fig = plot_with_trendline('hospital_beds_per_thousand','total_recovered_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('handwashing_facilities','reproduction_rate', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Percentage of population with basic handwashing facilities)')
fig.show(config=config)


#Function to create a correlation table

def create_corr_table(features1, features2):
    round_limit = 4
    features_total = features1 + features2
    header_vals = ['Feature','Feature2','r','r<sup>2</sup>','r (Avg df)','r<sup>2</sup> (Avg df)']
    r = [round(df[x].corr(df[y]),round_limit) for (x,y) in zip(features1,features2)]
    r2 = [round(x**2,round_limit) for x in r]
    r_avg = [round(df_avg_value[x].corr(df_avg_value[y]),round_limit) for (x,y) in zip(features1,features2)]
    r2_avg = [round(x**2,round_limit) for x in r_avg]
    
    def fix_name(name):
        name = "".join([(" "+i if i.isupper() else i) for i in name]).strip()
        name = name.replace('_',' ').title().replace('Hundred','100').replace('Thousand','1000')
        name = name.replace('Cardiovasc','Cardiovascular').replace('Gdp','GDP').replace('Gm ','')
        return name.replace('Death Percent','Case Fatality Percent').replace('Government','Govt.')
    
    def get_color_array(array):
        return [0 if abs(x)<=0.15 else 1 if abs(x)<=0.3 else 2 if abs(x)<=0.45 else 3 if abs(x)<=0.6 else 4 if 
                abs(x)<=0.75 else 5 for x in array]
    
    gen_colors = n_colors('rgb(247, 178, 178)','rgb(244, 40, 40)', 6, colortype='rgb')
    colors = [['lavender'] * len(features1)] + [['lavender'] * len(features2)]
    values_world = [[fix_name(x) for x in features1],
                    [fix_name(x) for x in features2],
                    r,
                    r2,
                    r_avg,
                    r2_avg]
    
    fig = go.Figure(data=[go.Table(columnwidth=[2,2,1,1,1,1],
    header=dict(values=header_vals,
                fill_color='paleturquoise',
                align='center',
                height=35,
                font=dict(size=15)),
    cells=dict(values=values_world,
               align='left',
               height=30,
               font=dict(size=13),
               fill_color=colors + [list(np.array(gen_colors)[get_color_array(x)]) for x in (r,r2,r_avg,r2_avg)]))
    ])
    height_padding = 140
    for (x,y) in zip(values_world[0],values_world[1]):
        if len(x) > 28:
            height_padding += 16.5
        if len(y) > 25:
            height_padding += 16.5
    
    fig.update_layout(title_text='Correlation Of Features',margin=dict(b=0), height=height_padding+len(features1)*33)
    fig.show(config=config)


features1 = ['total_deaths_per_hundred']*7 + ['Protection of elderly people','death_percent',
                                             'total_recovered_per_hundred','reproduction_rate']

features2 = ['median_age','aged_65_older','aged_70_older','diabetes_prevalence','cardiovasc_death_rate','male_smokers',
            'female_smokers','aged_65_older'] + ['hospital_beds_per_thousand']*2 + ['handwashing_facilities']

create_corr_table(features1, features2)


fig = plot_with_trendline('stringency_index','reproduction_rate',True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Value from 0 to 100; 100 = strictest response)')
fig.show(config=config)


fig = plot_with_trendline('stringency_index','percent_pop_affected', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(Value from 0 to 100; 100 = strictest response)')
fig.show(config=config)


fig = plot_with_trendline('Public information campaigns','total_vaccinations_per_hundred', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(0=No campaign, 2=Coordinated campaign across media)')
fig.show(config=config)


fig = plot_with_trendline('Emergency investment in healthcare','total_tests_per_hundred',True, log_x=True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(in USD)')
fig.show(config=config)


fig = plot_with_trendline('Investment in vaccines','total_vaccinations_per_hundred', True, log_x=True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(In USD)')
fig.show(config=config)


fig = plot_with_trendline('Vaccination policy','total_vaccinations_per_hundred', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(0=No availability, 5=Availability to all groups of population)')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[4,18,23]].columns, 'total_tests_per_hundred')
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.255, showarrow=False,
                   text='(CT- 0=No tracing, 2=Done for all cases<br>\
PIC- 0=No campaign, 2=Coordinated campaign across media<br>\
TP- 0=No policy, 3=Open public testing to asymptomatic people too)')
fig.update_layout(margin=dict(b=100))
fig.show(config=config)


fig = plot_with_trendline('Public information campaigns','reproduction_rate', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(0=No campaign, 2=Coordinated campaign across media)')
fig.show(config=config)


fig = plot_with_trendline('Facial Coverings', 'reproduction_rate', True)
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.17, showarrow=False,
                   text='(0=No policy, 4=Required outside home at all times)')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[3,21,22,25]].columns, 'reproduction_rate')
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.30, showarrow=False,
                   text='(CPT- 0=No measures, 2=Required Closing<br>\
SC- 0=No measures, 3=Required closing all levels<br>\
SAHR- 0=No measures, 3=Required not leaving house with minimal exceptions<br>\
WC- 0=No measures, 3=Required closing all but essential workplaces)')
fig.update_layout(margin=dict(b=120, t=20))
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[2,19,20]].columns, 'reproduction_rate')
fig.add_annotation(xref='paper', yref='paper', x=0.5,y=-0.25, showarrow=False,
                   text='(CPE- 0=No measures, 2=Required Canceling<br>\
ROG- 0=No restrictions, 4=Restrictions on gatherings of 10 people or less<br>\
ROIM- 0=No measures, 2=Movement restricted)')
fig.update_layout(margin=dict(b=100, t=20))
fig.show(config=config)


features1 = ['stringency_index']*2 + ['Public information campaigns','Emergency investment in healthcare',
                                     'Investment in vaccines','Vaccination policy','Contact tracing',
                                      'Public information campaigns','Testing policy','Public information campaigns',
                                     'Facial Coverings','Close public transport','School closing',
                                      'Stay at home requirements','Workplace closing','Cancel public events',
                                      'Restrictions on gatherings','Restrictions on internal movement']

features2 = ['reproduction_rate','percent_pop_affected','total_vaccinations_per_hundred','total_tests_per_hundred'] +\
             ['total_vaccinations_per_hundred']*2 + ['total_tests_per_hundred']*3 + ['reproduction_rate']*9

create_corr_table(features1, features2)


fig = plot_with_trendline('life_expectancy','death_percent', True)
fig.update_yaxes(title='Case Fatality Percent')
fig.show(config=config)


fig = plot_with_trendline('life_expectancy','recovered_percent', True)
fig.show(config=config)


fig = plot_with_trendline('life_expectancy','handwashing_facilities', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.18, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.22, showarrow=False,
                  text='(Handwashing facilities- Percentage of population with basic handwashing facilities)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


features1 = ['life_expectancy']*3

features2 = ['death_percent','recovered_percent','handwashing_facilities']

create_corr_table(features1, features2)


fig = plot_with_trendline('population_density', 'reproduction_rate', True, log_x=True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(sq. km)')
fig.show(config=config)


fig = plot_with_trendline('population_density', 'percent_pop_affected', True, log_x=True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(sq. km)')
fig.show(config=config)


fig = plot_with_trendline('population_density','stringency_index', True, log_x=True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(sq. km)')
fig.show(config=config)


features1 = ['population_density']*3

features2 = ['reproduction_rate','percent_pop_affected','stringency_index']

create_corr_table(features1, features2)


fig = plot_with_trendline('gdp_per_capita','hospital_beds_per_thousand', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','positive_rate', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Contact tracing', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.25, showarrow=False,
                  text='(CT- 0=No contact tracing, 2=Done for all cases)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','tests_per_case', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','total_tests_per_hundred', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','total_vaccinations_per_hundred', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[5,8,12]].columns, 'gdp_per_capita', True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Emergency investment in healthcare',True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.update_yaxes(title_text='Emergency Investment In Healthcare (USD)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Investment in vaccines',True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.update_yaxes(title_text='Investment In Vaccines (USD)')
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Fiscal measures',True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.update_yaxes(title_text='Fiscal Measures (USD)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Announced economic stimulus spending)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Income support',True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Income Support (for households)- 0=No income support, 2=Replacing 50% lost salary)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','Debt/contract relief',True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(D/C Relief (for households)- 0=No relief, 2=Broad debt/contract relief)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('gdp_per_capita','International support',True, log_y=True)
fig.add_annotation(xref='paper', yref='paper',x=0.5, y=-0.17, showarrow=False,
                  text='(2011 International$)')
fig.update_yaxes(title_text='International Support (USD)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(International Support means COVID-19 related aid to other countries)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


features1 = ['gdp_per_capita']*15

features2 = ['hospital_beds_per_thousand','positive_rate','Contact tracing','tests_per_case','total_tests_per_hundred',
            'total_vaccinations_per_hundred','ContainmentHealthIndex','EconomicSupportIndex','GovernmentResponseIndex',
             'Emergency investment in healthcare','Investment in vaccines','Fiscal measures','Income support',
             'Debt/contract relief','International support']

create_corr_table(features1, features2)


fig = plot_with_trendline('literacy_rate','reproduction_rate', True)
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','total_cases_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','death_percent', True)
fig.update_yaxes(title='Case Fatality Percent')
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','total_tests_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','total_vaccinations_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','handwashing_facilities', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Handwashing facilities- Percentage of population with basic handwashing facilities)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','hospital_beds_per_thousand', True)
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[5,8,12]].columns, 'literacy_rate', True)
fig.show(config=config)


fig = plot_with_trendline('literacy_rate','stringency_index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Stringency Index- Value from 0 to 100; 100 = strictest response)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


features1 = ['literacy_rate']*11

features2 = ['reproduction_rate','total_cases_per_hundred','death_percent','total_tests_per_hundred',
             'total_vaccinations_per_hundred','handwashing_facilities','hospital_beds_per_thousand',
             'ContainmentHealthIndex','EconomicSupportIndex','GovernmentResponseIndex','stringency_index']

create_corr_table(features1, features2)


fig = plot_with_trendline('Democracy index','reproduction_rate', True)
fig.show(config=config)


fig = plot_with_trendline('Democracy index','total_cases_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('Democracy index','death_percent', True)
fig.show(config=config)


fig = plot_with_trendline('Democracy index','recovered_percent', True)
fig.show(config=config)


fig = plot_with_trendline('Democracy index','total_tests_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('Democracy index','total_vaccinations_per_hundred', True)
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[2,3,20]].columns, 'Democracy index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(CPE- 0=No measures, 2=Require cancelling)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.31, showarrow=False,
                  text='(CPT- 0=No measures, 2=Require closing for most citizens)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.36, showarrow=False,
                  text='(ROIM- 0= No measures, 2= Restrictions in place)')
fig.update_layout(margin=dict(t=20,b=140))
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[21,22,25]].columns, 'Democracy index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='SC- 0=No measures, 3=Require closing all levels)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.31, showarrow=False,
                  text='(SAHR- 0=No measures, 3=Require not leaving house with minimal exceptions)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.36, showarrow=False,
                  text='(WC- 0=No measures, 3=Require closing except essential services)')
fig.update_layout(margin=dict(t=20,b=140))
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[10,15,19]].columns, 'Democracy index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(FC- 0=No policy, 4=Required at all times)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.31, showarrow=False,
                  text='(ITC- 0= No resctrictions, 4= Total border closure)')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.36, showarrow=False,
                  text='(ROG- 0= No restrictions, 4= Restrictions on gatherings of 10 people or less)')
fig.update_layout(margin=dict(t=20,b=140))
fig.show(config=config)


fig = plot_with_trendline('Democracy index','stringency_index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Stringency Index- Value from 0 to 100; 100 = strictest response)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


features1 = ['Democracy index']*16

features2 = ['reproduction_rate','total_cases_per_hundred','death_percent','recovered_percent',
             'total_tests_per_hundred','total_vaccinations_per_hundred','Cancel public events',
             'Close public transport','Restrictions on internal movement','School closing',
             'Stay at home requirements','Workplace closing','Facial Coverings','International travel controls',
             'Restrictions on gatherings','stringency_index']

create_corr_table(features1, features2)


fig = plot_with_trendline('human_development_index','total_tests_per_hundred', True)
fig.show(config=config)


fig = plot_with_trendline('human_development_index','total_vaccinations_per_hundred', True)
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.iloc[:,[5,8,12]].columns, 'human_development_index', True)
fig.show(config=config)


fig = plot_with_trendline('human_development_index','stringency_index', True)
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.21, showarrow=False,
                  text='For y-axis:')
fig.add_annotation(xref='paper', yref='paper',x=0, y=-0.26, showarrow=False,
                  text='(Stringency Index- Value from 0 to 100; 100 = strictest response)')
fig.update_layout(margin=dict(t=20,b=100))
fig.show(config=config)


features1 = ['human_development_index']*6

features2 = ['total_tests_per_hundred','total_vaccinations_per_hundred','ContainmentHealthIndex',
             'EconomicSupportIndex','GovernmentResponseIndex','stringency_index']

create_corr_table(features1, features2)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'total_deaths_per_hundred')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'Democracy index')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'stringency_index')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'literacy_rate')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'GovernmentResponseIndex')
fig.update_yaxes(title='Government Response Index')
fig.show(config=config)


fig = plot_multiple_trendlines(df_avg_value.columns[36:42], 'human_development_index')
fig.show(config=config)


features1 = ['gm_grocery_and_pharmacy','gm_parks', 'gm_residential','gm_retail_and_recreation','gm_transit_stations',
             'gm_workplaces']*6

features2 = ['total_deaths_per_hundred']*6 + ['Democracy index']*6 + ['stringency_index']*6 + ['literacy_rate']*6 +\
['GovernmentResponseIndex']*6 + ['human_development_index']*6

create_corr_table(features1, features2)


fig = px.scatter(df.groupby('country').max('reproduction_rate'), y='reproduction_rate',
           title='Highest Point of R<sub>0</sub> Per Country')

fig.update_traces(hovertemplate='<b>%{x}</b><br>R<sub>0</sub>: %{y}', marker_color='#db2b39')
fig.update_xaxes(title_text='')
fig.update_yaxes(title_text='')
fig.update_layout(margin=dict(l=30,r=10,t=70,b=0))
fig.show(config=config)


#Reordering dataframe to obtain correct order of date indices for animation frames

idx = df_25d[df_25d['country']=='United States'].index
temp_df = df_25d[df_25d['country']=='United States']
df_25d.drop(idx, inplace=True)
df_25d = pd.concat([temp_df, df_25d])

map_dates = list(temp_df.date.sort_values().dt.strftime('%d %b %Y')) * 22


fig = px.choropleth(df_25d, locations='iso_code', color='total_cases', animation_frame=df_25d.date.astype(str),
                        projection ='equirectangular', title='Total Cases',
                    color_continuous_scale=['#FCEEEF','#db2b39',"#7A151D"],
                   range_color=(0,np.max(df_25d.total_cases)), custom_data=[df_25d['country'],df_25d['total_cases']])

fig.update_layout(geo_showframe=False, margin=dict(r=0,l=0,t=50,b=0),height=700,
                  coloraxis_colorbar=dict(title=None,thickness=20,len=0.6))
fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]}')
fig.add_annotation(x=0.5,y=1,text=map_dates[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,.4s}')
    frame.layout.update(annotations=[dict(x=0.5,y=1,text=map_dates[idx],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


fig = px.choropleth(df_25d, locations='iso_code', color='total_deaths', animation_frame=df_25d.date.astype(str),
                        projection ='equirectangular', title='Total Deaths',
                    color_continuous_scale=['#DADAE7',"#3d405b",'#202031'],
                   range_color=(0,np.max(df_25d.total_deaths)), custom_data=[df_25d['country'],df_25d['total_deaths']])

fig.update_layout(geo_showframe=False, margin=dict(r=0,l=0,t=50,b=0),height=700,
                  coloraxis_colorbar=dict(title=None,thickness=20,len=0.6))
fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]}')
fig.add_annotation(x=0.5,y=1,text=map_dates[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,.4s}')
    frame.layout.update(annotations=[dict(x=0.5,y=1,text=map_dates[idx],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


#Filtering data to get desired dates to plot recoveries

df_25d_rec = df_25d.query('date < "08-05-2021"')
map_dates_rec = list(temp_df.query('date < "08-05-2021"').date.sort_values().dt.strftime('%d %b %Y')) * 9


fig = px.choropleth(df_25d_rec, locations='iso_code', color='total_recovered',
                    animation_frame=df_25d_rec.date.astype(str), projection ='equirectangular',title='Total Recovered',
                    color_continuous_scale=['#BDEFD3',"#2fbf71",'#186338'],
                   range_color=(0,np.max(df_25d_rec.total_recovered)),
                    custom_data=[df_25d_rec['country'],df_25d_rec['total_recovered']])

fig.update_layout(geo_showframe=False, margin=dict(r=0,l=0,t=50,b=0),height=700,
                  coloraxis_colorbar=dict(title=None,thickness=20,len=0.6))
fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]}')
fig.add_annotation(x=0.5,y=1,text=map_dates_rec[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,.4s}')
    frame.layout.update(annotations=[dict(x=0.5,y=1,text=map_dates_rec[idx],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


fig = px.choropleth(df_25d, locations='iso_code', color='total_tests', animation_frame=df_25d.date.astype(str),
                        projection ='equirectangular', title='Total Tests',
                    color_continuous_scale=['#FCD99C',"#faa613",'#9D6607'],
                   range_color=(0,np.max(df_25d.total_tests)), custom_data=[df_25d['country'],df_25d['total_tests']])

fig.update_layout(geo_showframe=False, margin=dict(r=0,l=0,t=50,b=0),height=700,
                  coloraxis_colorbar=dict(title=None,thickness=20,len=0.6))
fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]}')
fig.add_annotation(x=0.5,y=1,text=map_dates[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,.4s}')
    frame.layout.update(annotations=[dict(x=0.5,y=1,text=map_dates[idx],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


#Filtering data to get desired dates to plot vaccinations

df_25d_vax = df_25d.query('date > "27-10-2020"')
map_dates_vax = list(temp_df.query('date > "27-10-2020"').date.sort_values().dt.strftime('%d %b %Y')) * 9


fig = px.choropleth(df_25d_vax, locations='iso_code', color='total_vaccinations',
                    animation_frame=df_25d_vax.date.astype(str),
                    projection ='equirectangular', title='Total Vaccinations',
                    color_continuous_scale=['#D6F1FF',"#00a6fb",'#00527A'],
                    range_color=(0,np.max(df_25d.total_vaccinations)),
                    custom_data=[df_25d_vax['country'],df_25d_vax['total_vaccinations']])

fig.update_layout(geo_showframe=False, margin=dict(r=0,l=0,t=50,b=0),height=700,
                  coloraxis_colorbar=dict(title=None,thickness=20,len=0.6))
fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]}')
fig.add_annotation(x=0.5,y=1,text=map_dates_vax[0],showarrow=False, xref='paper',yref='paper',
                  font=dict(size=16))
for idx,frame in enumerate(fig.frames):
    frame.data[0].update(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,.4s}')
    frame.layout.update(annotations=[dict(x=0.5,y=1,text=map_dates_vax[idx],
                                           showarrow=False, xref='paper',yref='paper',font=dict(size=16))])
fig.show(config=config)


#Create dummy categories to group and plot as discrete colours

def create_dummy_feature(df, col):
    if col == 'total_deaths_per_hundred':
        max_value = round(max(df[col]),1)
        df['dummy'] = pd.cut(df[col], np.linspace(0,max_value,6))
    else:
        max_value = np.ceil(max(df[col]))
        df['dummy'] = pd.cut(df[col], np.linspace(0,max_value,6, dtype=int))
    df['dummy'] = df.dummy.astype(str).replace(({',':' to'}), regex=True)


#Checking for token
if mapbox_personal_token:
    px.set_mapbox_access_token(mapbox_personal_token)
    
    
#Function to create map with specifications
def create_mapbox(df, col):
    create_dummy_feature(df, col)
    fig = px.scatter_mapbox(df, lat='latitude', lon='longitude', size=col, size_max=20, zoom=1,
                            mapbox_style=('dark' if mapbox_personal_token else 'carto-darkmatter'),
                            title=col.replace('_',' ').title()+' Population', opacity=0.5,color='dummy',
                            custom_data=[df['country'],df[col]],
                            color_discrete_sequence=["#ff2c55","#34e4ea","#29bf12","#e3ff47","#c96bdb"])
#                   animation_frame=df.date.astype(str))
    fig.update_layout(margin=dict(l=0,r=0,t=80,b=0), legend=dict(title=None, orientation='h',
                                                                borderwidth=0))
    fig.update_traces(hovertemplate='<b>%{customdata[0]}</b><br>%{customdata[1]:,}<extra></extra>')
    fig.show(config=config)


create_mapbox(temp_df_cases, 'total_cases_per_hundred')


create_mapbox(temp_df_deaths, 'total_deaths_per_hundred')


create_mapbox(temp_df_recovered, 'total_recovered_per_hundred')


create_mapbox(temp_df_tests, 'total_tests_per_hundred')


create_mapbox(temp_df_vax, 'total_vaccinations_per_hundred')


#Reverting current theme to previous state

set_theme()


#Function to plot Top 10 countries of any feature

def plot_top10(df, feature):
    top10_countries = df.groupby('country').last()[feature].sort_values(ascending=False)[:10].index
    df = df.groupby(['country','date']).last().loc[top10_countries].reset_index()
    fig = px.line(df, x='date', y=feature, color='country', title='Top 10 '+feature.replace('_',' ').title(),
                 color_discrete_sequence= ["#db2b39","#3d405b","#2fbf71","#faa613","#00a6fb","#c04cfd","#f3de2c",
                                           "#735f3d","#6a788a","#054a29"])
    fig.update_layout(legend=dict(title=None))
    fig.update_traces(hovertemplate='%{x}<br><b>%{y}</b>')
    fig.show(config=config)


plot_top10(df, 'total_cases')


plot_top10(df, 'total_deaths')


plot_top10(df.query('date < "08-05-2021"'), 'total_recovered')


plot_top10(df, 'total_tests')


plot_top10(df.query('date > "12-1-2020"'), 'total_vaccinations')


#Picking 6 countries to plot with

countries_to_plot = ['Brazil','India','United States','Mauritius','New Zealand','Taiwan']
df_countries6 = pd.concat(df[df['country'] == country] for country in countries_to_plot)


#Preparing dataframe for parallel categories plot

df_oxbsg= df_countries6[['country','School closing','Workplace closing','Cancel public events',
                         'Restrictions on gatherings','Close public transport','Stay at home requirements',
                         'Restrictions on internal movement','International travel controls',
                         'Public information campaigns']]

df_oxbsg['country2'] = df_oxbsg['country']
df_oxbsg.replace({'country2':{'Brazil':1,'India':2,'United States':3,'Mauritius':4,'New Zealand':5,'Taiwan':6}},
                 inplace=True)


#Creating dimensions for parallel categories plot

dim = []
countries6_colorscale = ["#ff595e","#ffca3a","#8ac926","#1982c4","#6a4c93","#615055"]

for idx,var in enumerate(['country','Stay at home requirements','School closing','Workplace closing',
                          'Close public transport','Cancel public events','Restrictions on gatherings',
                          'Restrictions on internal movement','International travel controls',
                          'Public information campaigns']):
    dim.append(go.parcats.Dimension(values=df_oxbsg[var], label=var.title(), categoryorder='category descending'))


dim[0].categoryorder = 'array'
dim[0].categoryarray = ['Brazil','India','United States','Mauritius','New Zealand','Taiwan']

fig = go.Figure(data = [go.Parcats(dimensions=[x for x in dim],
        line={'color': df_oxbsg.country2,
              'colorscale':countries6_colorscale})])

fig.update_traces(hovertemplate='%{category}', labelfont_size=8,)
fig.update_layout(margin=dict(t=80,b=10,r=30,l=40), title='Government Measures')
fig.show(config=config)


#Preparing dimensions for parallel coordinates plot

df_countries6 = pd.concat(df[df['country'] == country] for country in countries_to_plot)
temp_df = pd.DataFrame({'country':df_countries6['country'].unique()})
temp_df['dummy'] = temp_df.index
df_countries6 = pd.merge(df_countries6, temp_df, on = 'country', how='left')

dimensions = list([dict(range=[0,df_countries6['dummy'].max()],
                       tickvals = temp_df['dummy'], ticktext = temp_df['country'],
                       label='Country', values=df_countries6['dummy']),
                  dict(range=[df_countries6['gm_retail_and_recreation'].min(),df_countries6['gm_retail_and_recreation'].max()],
                        label='Retail & Recreation', values=df_countries6['gm_retail_and_recreation']),
                  dict(range=[df_countries6['gm_grocery_and_pharmacy'].min(),df_countries6['gm_grocery_and_pharmacy'].max()],
                       label='Grocery & Pharmacy', values=df_countries6['gm_grocery_and_pharmacy']),
                  dict(range=[df_countries6['gm_parks'].min(),df_countries6['gm_parks'].max()],
                       label='Parks', values=df_countries6['gm_parks']),
                  dict(range=[df_countries6['gm_transit_stations'].min(),df_countries6['gm_transit_stations'].max()],
                       label='Transit Stations', values=df_countries6['gm_transit_stations']),
                  dict(range=[df_countries6['gm_workplaces'].min(),df_countries6['gm_workplaces'].max()],
                       label='Workplaces', values=df_countries6['gm_workplaces']),
                  dict(range=[df_countries6['gm_residential'].min(),df_countries6['gm_residential'].max()],
                       label='Residential', values=df_countries6['gm_residential']),
                  dict(range=[df_countries6['reproduction_rate'].min(),df_countries6['reproduction_rate'].max()],
                       label='Reproduction Rate', values=df_countries6['reproduction_rate'])
                  ])


countries6_colorscale = ["#ff595e","#ffca3a","#8ac926","#1982c4","#6a4c93","#003844"]
fig = go.Figure(data=go.Parcoords(line = dict(color = df_countries6['dummy'],
                                              colorscale = countries6_colorscale), dimensions=dimensions))

fig.update_layout(title='Mobility Trends')
fig.show(config=config)


#Modifying default plotting options for further graphs

pio.templates['my_theme'].layout.xaxis.title = {'standoff': 15}
pio.templates['my_theme'].layout.yaxis.title = {'standoff': 15}
pio.templates['my_theme'].layout.xaxis.tickformat = ''
pio.templates['my_theme'].layout.xaxis.hoverformat = ''


#Creating and displaying Benford's 1st digit's distribution

benford_dist1 = [np.log10(1+1/x)*100 for x in np.arange(1,10)]
fig = px.bar(x=np.arange(1,10), y=benford_dist1, title='Benford\'s 1st Digit Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency')
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's 2nd digit's distribution

benford_dist2 = [sum([np.log10(1+1/x)*100 for x in range(int('1'+str(y)),100,10)]) for y in np.arange(0,10)]
fig = px.bar(x=np.arange(0,10), y=benford_dist2, title='Benford\'s 2nd Digit Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=20)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's 3rd digit's distribution

benford_dist3 = [sum([np.log10(1+1/x)*100 for x in range(int('10'+str(y)),1000,10)]) for y in np.arange(0,10)]
fig = px.bar(x=np.arange(0,10), y=benford_dist3, title='Benford\'s 3rd Digit Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=20)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's 4th digit's distribution

benford_dist4 = [sum([np.log10(1+1/x)*100 for x in range(int('100'+str(y)),10000,10)]) for y in np.arange(0,10)]
fig = px.bar(x=np.arange(0,10), y=benford_dist4, title='Benford\'s 4th Digit Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=20)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's 5th digit's distribution

benford_dist5 = [sum([np.log10(1+1/x)*100 for x in range(int('1000'+str(y)),100000,10)]) for y in np.arange(0,10)]
fig = px.bar(x=np.arange(0,10), y=benford_dist5, title='Benford\'s 5th Digit Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=20)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's First 2 digits distribution

benford_dist_first_two = [np.log10(1+1/x)*100 for x in np.arange(10,100)]
fig = px.bar(x=np.arange(10,100), y=benford_dist_first_two, title='Benford\'s First 2 Digits Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=25)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's First 3 digits distribution

benford_dist_first_three = [np.log10(1+1/x)*100 for x in np.arange(100,1000)]
fig = px.bar(x=np.arange(100,1000), y=benford_dist_first_three, title='Benford\'s First 3 Digits Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=25)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Creating and displaying Benford's Last 2 digits distribution

benford_dist_last_two = [1.0] * 100
fig = px.bar(x=np.arange(0,100), y=benford_dist_last_two, title='Benford\'s Last 2 Digits Distribution')
fig.update_layout(xaxis_title='Digits', yaxis_title='Frequency', xaxis_nticks=25)
fig.update_traces(hovertemplate='Digit:<b>%{x}</b><br>Frequency:<b>%{y}</b>')
fig.show(config=config)


#Custom class to compute various digit frequencies of any series, along with its goodness of fit chi-square statistic 
#and p-value

class BenfordCompute(object):
    def __init__(self, series):
        self.series = series
    
    #Method to calculate frequency of single digits, 1st to 5th place
    def freq_single(self, order='first'):
        observed_freq = [0,0,0,0,0,0,0,0,0,0]
        for x in self.series.dropna().astype(int).values:
            x = abs(x)
            if order == 'first':
                digit = int(str(x)[0])
            try:
                check_variable = True
                if order == 'second':
                    digit = int(str(x)[1])
                elif order == 'third':
                    digit = int(str(x)[2])
                elif order == 'fourth':
                    digit = int(str(x)[3])
                elif order == 'fifth':
                    digit = int(str(x)[4])
            except IndexError:
                check_variable = False
                pass
            if check_variable:
                observed_freq[digit] += 1
        if order == 'first':
            del observed_freq[0]
        observed_freq = np.array(observed_freq)
        return observed_freq/observed_freq.sum()*100
    
    #Method to calculate frequency of first two digits
    def freq_first_two(self):
        observed_freq = [0]*90
        for x in self.series.dropna().astype(int).values:
            x = abs(x)
            if x < 10:
                continue
            firsttwo_digits = int(str(x)[:2])
            observed_freq[firsttwo_digits-10] += 1
        observed_freq = np.array(observed_freq)
        return observed_freq/observed_freq.sum()*100
    
    #Method to calculate frequency of last two digits
    def freq_last_two(self):
        observed_freq = [0]*100
        for x in self.series.dropna().astype(int).values:
            x = abs(x)
            if x < 10:
                continue
            lasttwo_digits = int(str(x)[-2:])
            observed_freq[lasttwo_digits] += 1
        observed_freq = np.array(observed_freq)
        return observed_freq/observed_freq.sum()*100
    
    #Method to calculate frequency of first three digits
    def freq_first_three(self):
        observed_freq = [0]*900
        for x in self.series.dropna().astype(int).values:
            x = abs(x)
            if x < 100:
                continue
            firstthree_digits = int(str(x)[:3])
            observed_freq[firstthree_digits-100] += 1
        observed_freq = np.array(observed_freq)
        return observed_freq/observed_freq.sum()*100
    
    #Method to calculate chi-square statistic and p-value for any digit frequencies
    def chi_square(self, order='first'):
        if '_' not in order:
            if order == 'first':
                expected_dist = benford_dist1
            elif order == 'second':
                expected_dist = benford_dist2
            elif order == 'third':
                expected_dist = benford_dist3
            elif order == 'fourth':
                expected_dist = benford_dist4
            elif order == 'fifth':
                expected_dist = benford_dist5
            return chisquare(self.freq_single(order=order), expected_dist)
        elif order == 'first_two':
            expected_dist = benford_dist_first_two
            return chisquare(self.freq_first_two(), expected_dist)
        elif order == 'last_two':
            expected_dist = benford_dist_last_two
            return chisquare(self.freq_last_two(), expected_dist)
        elif order == 'first_three':
            expected_dist = benford_dist_first_three
            return chisquare(self.freq_first_three(), expected_dist)
        
    def chi_square_pretty(self, order='first'):
        chi_pretty = self.chi_square(order=order)
        print('For '+order.title()+ ' Digit(s):\n')
        print('Chi-Square Statistic: ',chi_pretty[0])
        print('P-value: ',chi_pretty[1])
        
    #Method to calculate chi-square statistic and p-value for all digits together, presented in a table
    def chi_square_all(self, non_table=None):
        a = []
        to_tabulate = []
        a.append(self.chi_square(order='first'))
        a.append(self.chi_square(order='second'))
        a.append(self.chi_square(order='third'))
        a.append(self.chi_square(order='fourth'))
        a.append(self.chi_square(order='fifth'))
        a.append(self.chi_square(order='first_two'))
        a.append(self.chi_square(order='last_two'))
        a.append(self.chi_square(order='first_three'))
        for x,y,z in zip(['First Digit','Second Digit','Third Digit','Fourth Digit','Fifth Digit','First Two Digits',
                         'Last Two Digits','First Three Digits'],[x[0] for x in a],[x[1] for x in a]):
            to_tabulate.append([x,y,z])
        if non_table:
            return to_tabulate
        print(tabulate(to_tabulate,headers=['Digits','Chi-Square Statistic','p-value'], tablefmt='github'))


#Custom child class to plot figures for comparison of frequencies with Benford's distribution

class BenfordCheck(BenfordCompute):
    
    #Plot graph for frequency of single digits, 1st to 5th
    def fig_single(self, order='first'):
        start = 0
        fig = go.Figure()
        if order == 'first':
            start = 1
            expected_dist = benford_dist1
        elif order == 'second':
            expected_dist = benford_dist2
        elif order == 'third':
            expected_dist = benford_dist3
        elif order == 'fourth':
            expected_dist = benford_dist4
        elif order == 'fifth':
            expected_dist = benford_dist5
        pvalue = self.chi_square(order=order)[1]
        fig.add_bar(x=np.arange(start,10), y=self.freq_single(order=order), legendgroup='one',
                   hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.add_scatter(x=np.arange(start,10), y=expected_dist, legendgroup='two',
                       hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.update_layout(title_text='<b>'+order.title()+' Digit </b>(Observed vs Benford\'s) Frequencies\
<br><sup><span style="font-size:14px">p-value: '+str(pvalue)+'</span></sup>', legend_orientation='h',
                          legend_borderwidth=0)
        custom_legend_name(['Observed','Benford\'s'], fig_in_func=fig)
        return fig
    
    #Plot graph for frequency of first two digits
    def fig_first_two(self):
        fig = go.Figure()
        pvalue = self.chi_square(order='first_two')[1]
        fig.add_bar(x=np.arange(10,100), y=self.freq_first_two(), legendgroup='one',
                   hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.add_scatter(x=np.arange(10,100), y=benford_dist_first_two, legendgroup='two',
                       hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.update_layout(title_text='<b>First Two Digits </b>(Observed vs Benford\'s) Frequencies\
<br><sup><span style="font-size:14px">p-value: '+str(pvalue)+'</span></sup>', legend_orientation='h',
                          legend_borderwidth=0)
        custom_legend_name(['Observed','Benford\'s'], fig_in_func=fig)
        return fig
    
    #Plot graph for frequency of last two digits
    def fig_last_two(self):
        fig = go.Figure()
        pvalue = self.chi_square(order='last_two')[1]
        fig.add_bar(x=np.arange(0,100), y=self.freq_last_two(), legendgroup='one',
                   hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.add_scatter(x=np.arange(0,100), y=benford_dist_last_two, legendgroup='two',
                       hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.update_layout(title_text='<b>Last Two Digits </b>(Observed vs Benford\'s) Frequencies\
<br><sup><span style="font-size:14px">p-value: '+str(pvalue)+'</span></sup>', legend_orientation='h',
                         legend_borderwidth=0)
        custom_legend_name(['Observed','Benford\'s'], fig_in_func=fig)
        return fig
    
    #Plot graph for frequency of first three digits
    def fig_first_three(self):
        fig = go.Figure()
        pvalue = self.chi_square(order='first_three')[1]
        fig.add_bar(x=np.arange(100,1000), y=self.freq_first_three(), legendgroup='one',
                   hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.add_scatter(x=np.arange(100,1000), y=benford_dist_first_three, legendgroup='two',
                       hovertemplate='<b>Digit:</b> %{x}<br><b>Frequency:</b> %{y}<extra></extra>')
        fig.update_layout(title_text='<b>First Three Digits </b>(Observed vs Benford\'s) Frequencies\
<br><sup><span style="font-size:14px">p-value: '+str(pvalue)+'</span></sup>', legend_orientation='h',
                          legend_borderwidth=0)
        custom_legend_name(['Observed','Benford\'s'], fig_in_func=fig)
        return fig
    
    #Plot a single graph containing all the other plots as subplots
    def fig_all(self):
        pfirst = self.chi_square(order='first')[1]
        psecond = self.chi_square(order='second')[1]
        pthird = self.chi_square(order='third')[1]
        pfourth = self.chi_square(order='fourth')[1]
        pfifth = self.chi_square(order='fifth')[1]
        pfirsttwo = self.chi_square(order='first_two')[1]
        plasttwo = self.chi_square(order='last_two')[1]
        pfirstthree = self.chi_square(order='first_three')[1]
        fig = make_subplots(rows=2, cols=4, subplot_titles=['First Digit'+'<br><sup>p='+str(pfirst)+'</sup>',
                                                            'Second Digit'+'<br><sup>p='+str(psecond)+'</sup>',
                                                            'Third Digit'+'<br><sup>p='+str(pthird)+'</sup>',
                                                            'Fourth Digit'+'<br><sup>p='+str(pfourth)+'</sup>',
                                                            'Fifth Digit'+'<br><sup>p='+str(pfifth)+'</sup>',
                                                            'First Two Digits'+'<br><sup>p='+str(pfirsttwo)+'</sup>',
                                                            'Last Two Digits'+'<br><sup>p='+str(plasttwo)+'</sup>',
                                                        'First Three Digits'+'<br><sup>p='+str(pfirstthree)+'</sup>'])
        fig_11 = self.fig_single(order='first')
        fig_12 = self.fig_single(order='second')
        fig_13 = self.fig_single(order='third')
        fig_14 = self.fig_single(order='fourth')
        fig_21 = self.fig_single(order='fifth')
        fig_22 = self.fig_first_two()
        fig_23 = self.fig_last_two()
        fig_24 = self.fig_first_three()
        fig.append_trace(fig_11.data[0], row=1, col=1)
        fig.append_trace(fig_11.data[1], row=1, col=1)
        fig.append_trace(fig_12.data[0], row=1, col=2)
        fig.append_trace(fig_12.data[1], row=1, col=2)
        fig.append_trace(fig_13.data[0], row=1, col=3)
        fig.append_trace(fig_13.data[1], row=1, col=3)
        fig.append_trace(fig_14.data[0], row=1, col=4)
        fig.append_trace(fig_14.data[1], row=1, col=4)
        fig.append_trace(fig_21.data[0], row=2, col=1)
        fig.append_trace(fig_21.data[1], row=2, col=1)
        fig.append_trace(fig_22.data[0], row=2, col=2)
        fig.append_trace(fig_22.data[1], row=2, col=2)
        fig.append_trace(fig_23.data[0], row=2, col=3)
        fig.append_trace(fig_23.data[1], row=2, col=3)
        fig.append_trace(fig_24.data[0], row=2, col=4)
        fig.append_trace(fig_24.data[1], row=2, col=4)
        for trace in range(2,16):
            fig['data'][trace]['showlegend'] = False
        fig.update_layout(title_text='Benford\'s Law vs Observed Distribution', height=700, margin=dict(l=0,r=0),
                         colorway=['#DB2B39','#3D405B'], legend_x=0.9, legend_y=1.25, legend_borderwidth=1)
        return fig.show(config=config)


benford_all = BenfordCheck(df.new_cases)


benford_all.fig_single(order='first')


benford_all.chi_square_pretty(order='first')

For First Digit(s):

Chi-Square Statistic:  0.1309157275068432
P-value:  0.9999992740301535


benford_all.fig_single(order='second')


benford_all.chi_square_pretty(order='second')

For Second Digit(s):

Chi-Square Statistic:  0.02873508725030449
P-value:  0.9999999999035593


benford_all.fig_single(order='third')


benford_all.chi_square_pretty(order='third')

For Third Digit(s):

Chi-Square Statistic:  0.007681191212432483
P-value:  0.9999999999997432


benford_all.fig_single(order='fourth')


benford_all.chi_square_pretty(order='fourth')

For Fourth Digit(s):

Chi-Square Statistic:  0.05475238223200497
P-value:  0.9999999982638073


benford_all.fig_single(order='fifth')


benford_all.chi_square_pretty(order='fifth')

For Fifth Digit(s):

Chi-Square Statistic:  0.3514887377823567
P-value:  0.9999933806236577


benford_all.fig_first_two()


benford_all.chi_square_pretty(order='first_two')

For First_Two Digit(s):

Chi-Square Statistic:  0.6123103122725784
P-value:  1.0


benford_all.fig_first_three()


benford_all.chi_square_pretty(order='first_three')

For First_Three Digit(s):

Chi-Square Statistic:  3.544094272444643
P-value:  1.0


benford_all.fig_last_two()


benford_all.chi_square_pretty(order='last_two')

For Last_Two Digit(s):

Chi-Square Statistic:  5.642070969309949
P-value:  1.0


benford_saudi = BenfordCheck(df.query("country == 'Saudi Arabia'").new_cases)
benford_saudi.fig_all()


benford_saudi.chi_square_all()

| Digits             |   Chi-Square Statistic |      p-value |
|--------------------|------------------------|--------------|
| First Digit        |               21.0536  |   0.00700643 |
| Second Digit       |                3.57811 |   0.936927   |
| Third Digit        |                1.30022 |   0.998375   |
| Fourth Digit       |               10.3989  |   0.31917    |
| Fifth Digit        |              nan       | nan          |
| First Two Digits   |               38.2972  |   0.999999   |
| Last Two Digits    |               17.4933  |   1          |
| First Three Digits |              165.361   |   1          |


benford_iran = BenfordCheck(df.query("country == 'Iran'").new_cases)
benford_iran.fig_all()


benford_iran.chi_square_all()

| Digits             |   Chi-Square Statistic |   p-value |
|--------------------|------------------------|-----------|
| First Digit        |               18.7461  |  0.016278 |
| Second Digit       |                3.66679 |  0.931945 |
| Third Digit        |                1.64065 |  0.995963 |
| Fourth Digit       |                1.61582 |  0.996193 |
| Fifth Digit        |               11.6946  |  0.23108  |
| First Two Digits   |               43.0425  |  0.99999  |
| Last Two Digits    |               26.4815  |  1        |
| First Three Digits |              188.909   |  1        |


benford_usa = BenfordCheck(df.query("country == 'United States'").new_cases)
benford_usa.fig_all()


benford_usa.chi_square_all()

| Digits             |   Chi-Square Statistic |   p-value |
|--------------------|------------------------|-----------|
| First Digit        |                7.6367  |  0.469738 |
| Second Digit       |                2.30606 |  0.985653 |
| Third Digit        |                2.25802 |  0.9867   |
| Fourth Digit       |                1.02507 |  0.999378 |
| Fifth Digit        |                1.63354 |  0.99603  |
| First Two Digits   |               25.0341  |  1        |
| Last Two Digits    |               17.7394  |  1        |
| First Three Digits |              172.278   |  1        |


benford_india = BenfordCheck(df.query("country == 'India'").new_cases)
benford_india.fig_all()


benford_india.chi_square_all()

| Digits             |   Chi-Square Statistic |   p-value |
|--------------------|------------------------|-----------|
| First Digit        |                6.28408 |  0.615444 |
| Second Digit       |                3.57177 |  0.937275 |
| Third Digit        |                0.6059  |  0.999931 |
| Fourth Digit       |                2.13786 |  0.989095 |
| Fifth Digit        |                5.30009 |  0.807403 |
| First Two Digits   |               27.2529  |  1        |
| Last Two Digits    |               24.4231  |  1        |
| First Three Digits |              193.586   |  1        |


benford_china = BenfordCheck(df.query("country == 'China'").new_cases)
benford_china.fig_all()


benford_china.chi_square_all()

| Digits             |   Chi-Square Statistic |      p-value |
|--------------------|------------------------|--------------|
| First Digit        |                7.19379 | 0.515876     |
| Second Digit       |                3.99011 | 0.912062     |
| Third Digit        |               10.8287  | 0.287635     |
| Fourth Digit       |               15.6858  | 0.0737393    |
| Fifth Digit        |              899.941   | 6.36935e-188 |
| First Two Digits   |               37.6221  | 1            |
| Last Two Digits    |              184.817   | 3.80056e-07  |
| First Three Digits |             1046.74    | 0.000441485  |


benford_qatar = BenfordCheck(df.query("country == 'Qatar'").new_cases)
benford_qatar.fig_all()


benford_qatar.chi_square_all()

| Digits             |   Chi-Square Statistic |     p-value |
|--------------------|------------------------|-------------|
| First Digit        |               18.401   |   0.0184128 |
| Second Digit       |                2.37272 |   0.984113  |
| Third Digit        |                3.44786 |   0.94387   |
| Fourth Digit       |               18.8735  |   0.0262899 |
| Fifth Digit        |              nan       | nan         |
| First Two Digits   |               45.8099  |   0.999959  |
| Last Two Digits    |               20.4845  |   1         |
| First Three Digits |              176.724   |   1         |


benford_singapore = BenfordCheck(df.query("country == 'Singapore'").new_cases)
benford_singapore.fig_all()


benford_singapore.chi_square_all()

| Digits             |   Chi-Square Statistic |       p-value |
|--------------------|------------------------|---------------|
| First Digit        |                1.23894 |   0.996242    |
| Second Digit       |                2.2458  |   0.986958    |
| Third Digit        |                9.69554 |   0.375691    |
| Fourth Digit       |              275.123   |   4.87604e-54 |
| Fifth Digit        |              nan       | nan           |
| First Two Digits   |               23.495   |   1           |
| Last Two Digits    |              101.395   |   0.414445    |
| First Three Digits |              585.064   |   1           |


#Creatng dataframe for plotting of p-values of all countries

p_value_all = {}
for x in df.country.unique():
    benford_x = BenfordCheck(df[df['country'] == x].new_cases)
    chi_data = benford_x.chi_square_all(non_table=True)
    p_data = [x[2] for x in chi_data]
    if np.isnan(np.array(p_data)).all():
        continue
    p_value_all[x] = p_data
p_value_df = pd.DataFrame(p_value_all).T


#Function that helps in filling colours in table

def p_value_color(array):
    return [0 if np.isnan(x) else 1 if x>=0.05 else 2 if x<0.05 else None for x in array]


fig = go.Figure(data=[go.Table(#columnwidth=[200,100,100,100,100,100,100,100,100],
    header=dict(values=['Countries','First Digit','Second Digit','Third Digit','Fourth Digit','Fifth Digit',
                        'First Two Digits','Last Two Digits','First Three Digits'],
                fill_color='paleturquoise',
                align='left'),
    cells=dict(values=[p_value_df.index, round(p_value_df[0],5),round(p_value_df[1],5),round(p_value_df[2],5),
                       round(p_value_df[3],5),round(p_value_df[4],5),round(p_value_df[5],5),round(p_value_df[6],5),
                       round(p_value_df[7],5)],align='left',
              fill_color=[['lavender']*p_value_df.shape[0]] + [list(np.array(['white','#a2f562',
                                                                              '#f56262'])[p_value_color(x)]) for x 
                                    in (p_value_df[0], p_value_df[1], p_value_df[2], p_value_df[3], p_value_df[4],
                                        p_value_df[5], p_value_df[6], p_value_df[7])]))
])
fig.update_layout(title_text='P-Values Of All Countries',margin=dict(b=0,l=0,r=0), height=700)
fig.show(config=config)


# del df
# del df_gov
# del recovered_df
# del google_df
# del df_25d
# del df_25d_rec
# del df_25d_vax
# del df_countries6
# del continent_df
# del continent_df_rec
# del continent_df_vax
# del us_recovered_df
# del df_oxbsg
# del temp_continent_df
# del world_df_tests
# del world_df_rec
# del temp_df_vax
# del temp_df_cases
# del temp_df_deaths
# del temp_df_recovered
# del df_avg_value
# del world_df_vax
# del temp_df_tests
# del temp_df_values2
# del rec_df
# del continent_df_25d
# del continent_df_25d_rec
# del continent_df_25d_vax
# del temp_world_df
# del date_cols


import itertools
from pmdarima.arima import auto_arima
from prophet.diagnostics import cross_validation
from prophet.diagnostics import performance_metrics
from prophet import Prophet
from scipy.special import inv_boxcox
from scipy.stats import boxcox
from sklearn.linear_model import LinearRegression
from statsmodels.tsa.holtwinters import ExponentialSmoothing
from statsmodels.tsa.statespace.sarimax import SARIMAX
from statsmodels.tsa.stattools import acf
from statsmodels.tsa.stattools import adfuller
from statsmodels.tsa.stattools import kpss
from statsmodels.tsa.stattools import pacf
from tqdm import tqdm


from helper_functions import config
from helper_functions import set_theme
from helper_functions import custom_legend_name
world_df = pd.read_csv('final_world_df.csv')
world_df['date'] = pd.to_datetime(world_df['date'], format=('%Y-%m-%d'))


#Setting theme for further plots

set_theme()


#Making a dataframe with `new_cases` as the target variable

df = world_df
df = df[['date','new_cases']]
df['new_cases'] = df['new_cases'].fillna(df['new_cases'].bfill())
df = df.set_index('date')
df.index = pd.to_datetime(df.index)


fig = px.line(df.new_cases, title='New Cases')
fig.update_traces(showlegend=False, hovertemplate='%{x}<br><b>%{y:.5s}</b><extra></extra>')
fig.show(config=config)


#Function to plot moving average and standard deviation to visually inspect stationarity

def test_stationarity(df, feature, series_type='val'):
    rol_mean = df[feature].rolling(window=7).mean()
    rol_std = df[feature].rolling(window=7).std()
    fig = go.Figure()
    fig.add_scatter(y=df[feature], x=df.index)
    fig.add_scatter(y=rol_mean, x=df.index)
    fig.add_scatter(y=rol_std, x=df.index)
    fig.update_layout(legend=dict(title=None),
                  hovermode='x unified',hoverlabel=dict(bgcolor='rgba(255,255,255,0.75)',font=dict(color='black')))
    custom_legend_name(['New Cases','Rolling 7-Day Mean','Rolling 7-Day Standard Deviation'],fig_in_func=fig)
    if series_type == 'full':
        fig.update_layout(title_text='Full Non-Split Series')
    elif series_type == 'val':
        fig.update_layout(title_text='Training Set')
    elif series_type == 'val2':
        fig.update_layout(title_text='(Training + 1st Validation) Set')
    elif series_type == 'test':
        fig.update_layout(title_text='(Training + Val1 + Val2) Set')
    fig.show(config=config)


test_stationarity(df, 'new_cases', 'full')


#Function to output results of ADF test

def adf_test(timeseries):
    print("Results of Dickey-Fuller Test:")
    dftest = adfuller(timeseries)
    dfoutput = pd.Series(
        dftest[0:4],
        index=[
            "Test Statistic",
            "p-value",
            "#Lags Used",
            "Number of Observations Used",
        ],
    )
    for key, value in dftest[4].items():
        dfoutput["Critical Value (%s)" % key] = value
    print(dfoutput)


adf_test(df.new_cases)

Results of Dickey-Fuller Test:
Test Statistic                  -1.826814
p-value                          0.367233
#Lags Used                      14.000000
Number of Observations Used    592.000000
Critical Value (1%)             -3.441444
Critical Value (5%)             -2.866435
Critical Value (10%)            -2.569377
dtype: float64


#Function to output result of KPSS test

def kpss_test(timeseries):
    print("Results of KPSS Test:")
    kpsstest = kpss(timeseries, nlags="auto")
    kpss_output = pd.Series(
        kpsstest[0:3], index=["Test Statistic", "p-value", "Lags Used"]
    )
    for key, value in kpsstest[3].items():
        kpss_output["Critical Value (%s)" % key] = value
    print(kpss_output)


kpss_test(df.new_cases)

Results of KPSS Test:
Test Statistic            2.885272
p-value                   0.010000
Lags Used                16.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1907: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is smaller than the p-value returned.


#Separating data into sets of weeks

weeks = [g for n, g in df.groupby(pd.Grouper(freq='W'))]
fig = go.Figure()
for idx,x in enumerate(weeks):
    if idx >55:
        fig.add_scatter(x=x.index, y=x.new_cases)
fig.update_layout(title_text='Part Of Series Showing Weekly Seasonality', showlegend=False, yaxis_title='New Cases',
                 yaxis_title_font_color='#2A3F5F')
fig.update_traces(hovertemplate='%{x}<br><b>%{y:.5s}</b><extra></extra>')
fig.show(config=config)
for week in weeks:
    week.index = week.index.day_name()
(weeks[0],weeks[9]) = (weeks[9],weeks[0])


fig = go.Figure()
fig.update_layout(showlegend=False, title_text='Plots Showing Weekly Seasonality For Each Week<br>\
Since the Start of the Pandemic')
for idx,x in enumerate(weeks):
    if idx%10 == 0 and idx>0:
        fig.show(config=config)
        fig = go.Figure()
        fig.update_layout(showlegend=False)
    fig.add_scatter(x=x.index, y=x.new_cases)
fig.show(config=config)


#Function to split dataframe into train, test, validation 1 and validation 2

def split_train_test_val(df):
    test_size = 30
    train_df = df[:-test_size*3]
    val_df = df[-test_size*3:-test_size*2]
    val2_df = df[-test_size*2:-test_size]
    test_df = df[-test_size:]
    return train_df, test_df, val_df, val2_df


#Spliting data

train_df, test_df, val_df, val2_df = split_train_test_val(df)


test_stationarity(train_df, 'new_cases')


#Function to create ACF or PACF plots

def create_corr_plot(series, plot_pacf=False):
    corr_array = pacf(series.dropna(), alpha=0.05) if plot_pacf else acf(series.dropna(), alpha=0.05)
    lower_y = corr_array[1][:,0] - corr_array[0]
    upper_y = corr_array[1][:,1] - corr_array[0]

    fig = go.Figure()
    [fig.add_scatter(x=(x,x), y=(0,corr_array[0][x]), mode='lines',line_color='#3f3f3f') 
     for x in range(len(corr_array[0]))]
    fig.add_scatter(x=np.arange(len(corr_array[0])), y=corr_array[0], mode='markers', marker_color='#1f77b4',
                   marker_size=12)
    fig.add_scatter(x=np.arange(len(corr_array[0])), y=upper_y, mode='lines', line_color='rgba(255,255,255,0)')
    fig.add_scatter(x=np.arange(len(corr_array[0])), y=lower_y, mode='lines',fillcolor='rgba(32, 146, 230,0.3)',
            fill='tonexty', line_color='rgba(255,255,255,0)')
    fig.update_traces(showlegend=False)
    fig.update_xaxes(range=[-1,42], tickformat='', hoverformat='')
    fig.update_yaxes(zerolinecolor='#000000')
    fig.update_traces(hovertemplate='%{x}<br>%{y}<extra></extra>')
    
    title='Partial Autocorrelation (PACF)' if plot_pacf else 'Autocorrelation (ACF)'
    fig.update_layout(title=title)
    fig.show(config=config)


create_corr_plot(train_df.new_cases)


create_corr_plot(train_df.new_cases, True)


#Function to set inferred index for df to prevent statsmodels throwing warnings

def set_inferred_index(df):
    try:
        df.index = pd.DatetimeIndex(df.index.values, freq=df.index.inferred_freq)
    except AttributeError:
        pass
    
set_inferred_index(train_df)


#Function to get actual values, predicted values and model in a tuple

def get_model_pred(model_arg, aic_bic=True, split_type='val', series_type='normal', algo_type='sarima'):
    predicted = []
    if split_type == 'test':
        model_df = train_df.append(val_df).append(val2_df)[['new_cases']].copy()
        future_df = test_df.copy()            
    elif split_type == 'val2':
        model_df = train_df.append(val_df)[['new_cases']].copy()
        future_df = val2_df.copy()
    else:
        model_df = train_df[['new_cases']].copy()
        future_df = val_df.copy()
    if algo_type == 'hw':
        if model_arg[0] == 'mul' or model_arg[1] == 'mul':
            model_df = model_df.drop(model_df[(model_df.new_cases == 0)].index)
    if series_type == 'lr':
        X = [i for i in range(len(model_df.new_cases))]
        X = np.reshape(X, (len(X), 1))
        lr = LinearRegression().fit(X,model_df.new_cases)
        lr_trend = lr.predict(X)
        model_df = model_df.new_cases - lr_trend
    elif series_type == 'boxcox':
        model_df,lmbda = boxcox(model_df.drop(model_df[(model_df.new_cases == 0)].index).new_cases, lmbda=None)
    pred_index = len(model_df)
    for x in range(len(future_df)):
        set_inferred_index(model_df)
        if algo_type == 'sarima':
            order = model_arg[0]
            seasonal_order = model_arg[1]
            model = SARIMAX(model_df, order=order, seasonal_order=seasonal_order)
            model = model.fit(maxiter=200)
            if aic_bic:
                return (order, seasonal_order),model.aic, model.bic
        elif algo_type == 'hw':
            model = ExponentialSmoothing(model_df, seasonal_periods=model_arg[7], seasonal=model_arg[1],
                                         trend=model_arg[0], damped=model_arg[2])
            model = model.fit(smoothing_level=model_arg[4], smoothing_trend=model_arg[5],
                              smoothing_seasonal=model_arg[6], damping_trend=model_arg[3])
            if aic_bic:
                return (model_arg),model.aic, model.bic
        if series_type == 'lr':
            inversed_forecast = model.forecast() + lr.predict([[pred_index]])
            predicted.append(inversed_forecast)
            detrended_future_obs = future_df[['new_cases']].iloc[x] - lr.predict([[pred_index]])
            model_df = model_df.append(detrended_future_obs)
            pred_index += 1
        elif series_type == 'boxcox':
            inversed_forecast = inv_boxcox(model.forecast(),lmbda)
            predicted.append(inversed_forecast)
            detrended_future_obs = ((future_df[['new_cases']].iloc[x])**lmbda - 1) / lmbda
            model_df = np.append(model_df, detrended_future_obs)
        else:
            predicted.append(model.forecast())
            model_df = model_df.append(future_df[['new_cases']].iloc[x])
    if series_type == 'lr':
        return future_df.new_cases, np.array([x.values[0] for x in predicted]), model
    return future_df.new_cases, np.array([x[0] for x in predicted]), model


#Function to plot actual and predicted values

def plot_pred(actual, predicted, plot_model_attribs):
    fig = go.Figure()
    fig.add_scatter(y=actual, x=actual.index)
    fig.add_scatter(y=predicted, x=actual.index)
    fig.update_xaxes(tickformat='', hoverformat='')
    custom_legend_name(['Actual','Predicted'], fig_in_func=fig)
    fig.update_layout(title_text=('Actual vs Predicted<br><span style="font-size:0.65em">'+
                                  str(plot_model_attribs)+'</span>'), legend_orientation='h',
                     legend_borderwidth=0)
    fig.update_traces(hovertemplate='%{x}<br><b>%{y:.5s}</b><extra></extra>')
    fig.show(config=config)


#Function to plot residuals

def plot_resid(actual, predicted, plot_model_attribs):
    """
    resid:              The y-axis points to be plotted. x-axis is assumed to be linearly increasing.
    tickvalues:         The values you want to be displayed as ticklabels on x-axis. Works for hoverlabels too. Has 
                        to be same length as `resid`. (This is necessary to ignore 'gaps' in graph where two polygons 
                        meet.)
    plot_model_attribs: The parameters passed to the SARIMA model.
    """
    #Adjusting array with paddings to connect polygons at zero line on x-axis
    resid = actual - predicted
    index_array = []
    start_digit = 0
    split_array = np.split(resid,np.where(np.abs(np.diff(np.sign(resid)))==2)[0]+1)
    split_array = [np.append(x,0) for x in split_array]
    split_array = [np.insert(x,0,0) for x in split_array]
    split_array[0] = np.delete(split_array[0],0)
    split_array[-1] = np.delete(split_array[-1],-1)
    for x in split_array:
        index_array.append(np.arange(start_digit,start_digit+len(x)))
        start_digit += len(x)-1

    #Formatting dates to be displayed neatly as autoformatting won't work for this figure
    datelist = []
    cur_month = ''
    prev_month = ''
    cur_year = ''
    for x in actual.index:
        ts = pd.to_datetime(x)
        ts = ts.strftime('%Y %b %d')
        if cur_month == ts[5:-3]:
            datelist.append(pd.to_datetime(x).strftime('%d'))
        elif cur_year == ts[:4]:
            datelist.append(pd.to_datetime(x).strftime('%b %d'))
            prev_month = cur_month
        else:
            datelist.append(ts)
        cur_month = ts[5:-3]
        cur_year = ts[:4]

    #Making an array for ticklabels
    flat = []
    for x in index_array:
        for y in x:
            flat.append(y)
    flat_counter = Counter(flat)
    none_indices = np.where([(flat_counter[x]>1) for x in flat_counter])[0]
    custom_tickdata = []
    neg_padding = 0
    start_pos = 0
    for y in range(len(flat)):
        for x in range(start_pos,flat[-1]+1):
            if x in none_indices:
                custom_tickdata.append('')
                break
            custom_tickdata.append(datelist[x-neg_padding])
        neg_padding +=1
        start_pos = 1+x

    #Making an array for hoverlabels
    custom_hoverdata=[]
    sublist = []
    len_track = 2
    value_idx = 0
    for idx,x in enumerate(custom_tickdata):
        if x == '':
            custom_hoverdata.append(sublist)
            sublist = []
            sublist.append(x)
            continue
        if idx==0:
            sublist.append(str(x)+'<br><b>'+str(resid[value_idx])+'</b>')
            value_idx += 1
        else:
            if len(x) > len_track:
                prev_month = cur_month
            if prev_month == '':
                sublist.append(str(x)+' '+cur_month+'<br><b>'+str(resid[value_idx])+'</b>')
                value_idx += 1
            else:
                sublist.append(str(x)+' '+prev_month+'<br><b>'+str(resid[value_idx])+'</b>')
                value_idx += 1
        sublist2 = sublist.copy()
    custom_hoverdata.append(sublist2)

    #Creating figure
    fig = go.Figure()
    idx = 0
    for x,y in zip(split_array,index_array):
        color = 'rgba(219,43,57,0.8)' if x[1]<0 else 'rgba(47,191,113,0.8)'
        if (idx==0 and x[0] < 0):
            color= 'rgba(219,43,57,0.8)'
        fig.add_scatter(y=x, x=y, fill='tozeroy', fillcolor=color, line_color=color, customdata=custom_hoverdata[idx],
                        hovertemplate='%{customdata}<extra></extra>')
        idx += 1
    fig.update_layout(margin=dict(l=0), title_text='Residuals<br><span style="font-size:0.65em">'+
                                  str(plot_model_attribs)+'</span>')
    fig.update_xaxes(hoverformat='',tickmode = 'array', tickangle=90,
                tickvals = np.arange(index_array[-1][-1]+1),
                ticktext = custom_tickdata)
    fig.update_traces(mode='lines', showlegend=False)
    fig.show(config=config)


#Class containing the model, figure and metrics of the predictions

class Metrics(object):
    def __init__(self, actual, predicted, model):
        self.actual = actual
        self.predicted = predicted
        self.model = model
    def mse(self):
        return np.sum(np.square(self.actual-self.predicted))/len(self.predicted)
    def rmse(self):
        return np.sqrt(self.mse())
    def mae(self):
        return np.sum(np.abs(self.actual-self.predicted))/len(self.predicted)
    def mape(self):
        try:
            return np.mean(np.abs((self.actual-self.predicted)/self.actual))*100
        except ZeroDivisionError:
            return np.nan
    def mase(self, algo_type='sarima'):
        try:
            if algo_type == 'hw':
                s = 7
            else:
                s = self.diff_order()[0] if not self.diff_order()[1] else self.diff_order()[1]
                if s == 0:
                    s = 1
            y_train = np.asarray(self.model.model.endog[:-29])
            mae_naive = np.mean(np.abs(y_train[s:] - y_train[:-s]))
            if mae_naive == 0:
                return np.nan
            else:
                return np.mean(np.abs(self.actual-self.predicted))/mae_naive
        except TypeError:
            return None
    def fig(self):
        return plot_pred(self.actual, self.predicted, self.model_attribs)
    def resid_fig(self):
        return plot_resid(self.actual, self.predicted, self.model_attribs)
    def model_name(self, model_attribs, seasonal_model_attribs):
        if seasonal_model_attribs == 0:
            self.model_attribs = (model_attribs)
        else:
            self.model_attribs = (model_attribs,seasonal_model_attribs)
    def aic(self):
        return self.model.aic
    def bic(self):
        return self.model.bic
    def diff_order(self):
        try:
            return (self.model.model.k_diff, self.model.model.k_seasonal_diff)
        except:
            pass
    def get_all(self, algo_type='sarima'):
        print('For model ' + str(self.model_attribs) + ':\n\nMean Squared Error (MSE): ',self.mse(),
             '\nRoot Mean Squared Error (RMSE): ', self.rmse(),
             '\nMean Absolute Error (MAE): ', self.mae(),
             '\nMean Absolute Percentage Error (MAPE): ', self.mape(),
             '\nMean Absolute Scaled Error (MASE): ', self.mase(algo_type))


#Standard deviation after single order of non-seasonal differencing

train_df['diff1'] = train_df['new_cases'] - train_df['new_cases'].shift(1)
train_df['diff1'].std()

51023.727210007055


#Standard deviation after another order of non-seasonal differencing

train_df['diff2'] = train_df['diff1'] - train_df['diff1'].shift(1)
train_df['diff2'].std()

65671.92907876163


test_stationarity(train_df, 'diff1')


adf_test(train_df.diff1.dropna())

Results of Dickey-Fuller Test:
Test Statistic                  -4.820362
p-value                          0.000050
#Lags Used                       8.000000
Number of Observations Used    507.000000
Critical Value (1%)             -3.443314
Critical Value (5%)             -2.867258
Critical Value (10%)            -2.569815
dtype: float64


kpss_test(train_df.diff1.dropna())

Results of KPSS Test:
Test Statistic            0.103961
p-value                   0.100000
Lags Used                39.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1911: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is greater than the p-value returned.


create_corr_plot(train_df.diff1)


create_corr_plot(train_df.diff1, True)


#Training ARIMA with I=1 and MA=2

sarima_diff1 = ((0,1,2),(0,0,0,0))
actual_diff1, predicted_diff1, model_diff1 = get_model_pred(sarima_diff1, aic_bic=False)
diff1 = Metrics(actual_diff1, predicted_diff1, model_diff1)
diff1.model_name(sarima_diff1[0], sarima_diff1[1])


diff1.fig()


diff1.resid_fig()


diff1.get_all()

For model ((0, 1, 2), (0, 0, 0, 0)):

Mean Squared Error (MSE):  1660829174.6651156 
Root Mean Squared Error (RMSE):  40753.27195042277 
Mean Absolute Error (MAE):  33540.486042811695 
Mean Absolute Percentage Error (MAPE):  9.24658872542414 
Mean Absolute Scaled Error (MASE):  1.0229865045016144


#Standard deviation after single order of seasonal differencing

train_df['sdiff'] = train_df['new_cases'] - train_df['new_cases'].shift(7)
train_df['sdiff'].std()

50242.043818916325


#Standard deviation after another order of seasonal differencing

train_df['sdiff2'] = train_df['sdiff'] - train_df['sdiff'].shift(7)
train_df['sdiff2'].std()

58529.80502841225


test_stationarity(train_df, 'sdiff')


adf_test(train_df.sdiff.dropna())

Results of Dickey-Fuller Test:
Test Statistic                  -2.845976
p-value                          0.052001
#Lags Used                      15.000000
Number of Observations Used    494.000000
Critical Value (1%)             -3.443657
Critical Value (5%)             -2.867408
Critical Value (10%)            -2.569896
dtype: float64


kpss_test(train_df.sdiff.dropna())

Results of KPSS Test:
Test Statistic            0.131353
p-value                   0.100000
Lags Used                12.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1911: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is greater than the p-value returned.


create_corr_plot(train_df.sdiff)


create_corr_plot(train_df.sdiff, True)


#Training SARIMA with S=7, S(AR)=1 and S(I)=1

sarima_sdiff = ((0,0,0),(1,1,0,7))
actual_sdiff, predicted_sdiff, model_sdiff = get_model_pred(sarima_sdiff, aic_bic=False)
sdiff = Metrics(actual_sdiff, predicted_sdiff, model_sdiff)
sdiff.model_name(sarima_sdiff[0], sarima_sdiff[1])


sdiff.fig()


sdiff.resid_fig()


sdiff.get_all()

For model ((0, 0, 0), (1, 1, 0, 7)):

Mean Squared Error (MSE):  1089155827.5593162 
Root Mean Squared Error (RMSE):  33002.36093917095 
Mean Absolute Error (MAE):  27454.057399432542 
Mean Absolute Percentage Error (MAPE):  7.02263806347358 
Mean Absolute Scaled Error (MASE):  0.8373501259815913


#Standard deviation after single order seasonal and non-seasonal difference

train_df['sdiff_plus_diff'] = train_df['sdiff'] - train_df['sdiff'].shift(1)
train_df['sdiff_plus_diff'].std()

38450.48712209421


test_stationarity(train_df,'sdiff_plus_diff')


adf_test(train_df.sdiff_plus_diff.dropna())

Results of Dickey-Fuller Test:
Test Statistic                -8.388521e+00
p-value                        2.401647e-13
#Lags Used                     1.400000e+01
Number of Observations Used    4.940000e+02
Critical Value (1%)           -3.443657e+00
Critical Value (5%)           -2.867408e+00
Critical Value (10%)          -2.569896e+00
dtype: float64


kpss_test(train_df.sdiff_plus_diff.dropna())

Results of KPSS Test:
Test Statistic            0.115054
p-value                   0.100000
Lags Used                78.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1911: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is greater than the p-value returned.


create_corr_plot(train_df.sdiff_plus_diff)


create_corr_plot(train_df.sdiff_plus_diff, True)


#Training SARIMA with AR=2, I=1, S=7, S(AR)=3 and S(I)=1

sarima_sdiff_plus_diff = ((2,1,0),(3,1,0,7))
actual_sdiff_plus_diff, predicted_sdiff_plus_diff, model_sdiff_plus_diff = get_model_pred(sarima_sdiff_plus_diff,
                                                                                          aic_bic=False)
sdiff_plus_diff = Metrics(actual_sdiff_plus_diff, predicted_sdiff_plus_diff, model_sdiff_plus_diff)
sdiff_plus_diff.model_name(sarima_sdiff_plus_diff[0],sarima_sdiff_plus_diff[1])


sdiff_plus_diff.fig()


sdiff_plus_diff.resid_fig()


sdiff_plus_diff.get_all()

For model ((2, 1, 0), (3, 1, 0, 7)):

Mean Squared Error (MSE):  553469160.9893118 
Root Mean Squared Error (RMSE):  23525.92529507207 
Mean Absolute Error (MAE):  19136.51106000688 
Mean Absolute Percentage Error (MAPE):  5.024600483361559 
Mean Absolute Scaled Error (MASE):  0.5836645459652926


#Transforming dataframe by boxcox to train a model on it

boxcox_df = train_df.drop(train_df[(train_df.new_cases == 0)].index)
series,lmbda = boxcox(boxcox_df.new_cases, lmbda=None)
boxcox_df['new_cases'] = series


fig = px.histogram(train_df.new_cases)
fig.layout.xaxis.tickformat = ''
fig.layout.xaxis.hoverformat = ''
fig.update_layout(legend_borderwidth=1, legend_title=None, title_text='Before Transformation')
fig.update_traces(hovertemplate='%{x}<br>%{y}<extra></extra>')
fig.show(config=config)


fig = px.histogram(boxcox_df.new_cases)
fig.layout.xaxis.tickformat = ''
fig.layout.xaxis.hoverformat = ''
fig.update_layout(legend_borderwidth=1, legend_title=None, title_text='After Transformation')
fig.update_traces(hovertemplate='%{x}<br>%{y}<extra></extra>')
fig.show(config=config)


test_stationarity(boxcox_df, 'new_cases')


lmbda

0.5125126023622883


adf_test(series)

Results of Dickey-Fuller Test:
Test Statistic                  -1.956145
p-value                          0.306127
#Lags Used                      15.000000
Number of Observations Used    479.000000
Critical Value (1%)             -3.444076
Critical Value (5%)             -2.867593
Critical Value (10%)            -2.569994
dtype: float64


kpss_test(series)

Results of KPSS Test:
Test Statistic            3.439288
p-value                   0.010000
Lags Used                12.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1907: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is smaller than the p-value returned.


#Finding standard deviation for four types of differencing on boxcox transformed series

boxcox_df['diff1'] = boxcox_df['new_cases'] - boxcox_df['new_cases'].shift(1)
boxcox_df['diff2'] = boxcox_df['diff1'] - boxcox_df['diff1'].shift(1)
boxcox_df['sdiff'] = boxcox_df['new_cases'] - boxcox_df['new_cases'].shift(7)
boxcox_df['sdiff_plus_diff'] = boxcox_df['diff1'] - boxcox_df['diff1'].shift(7)


boxcox_df['new_cases'].std()

589.9032948922065


boxcox_df['diff1'].std()

88.65287213815147


boxcox_df['diff2'].std()

118.11240422574501


boxcox_df['sdiff'].std()

90.02280779123264


boxcox_df['sdiff_plus_diff'].std()

72.23035547777096


create_corr_plot(boxcox_df.sdiff_plus_diff)


create_corr_plot(boxcox_df.sdiff_plus_diff, plot_pacf=True)


#Training SARIMA with AR=2, I=1, S=7, S(AR)=3 and S(I)=1 on boxcox transformed series

sarima_boxcox = ((2,1,0),(3,1,0,7))
actual_boxcox, predicted_boxcox, model_boxcox = get_model_pred(sarima_boxcox, aic_bic=False, series_type='boxcox')
boxcox_trained = Metrics(actual_boxcox, predicted_boxcox, model_boxcox)
boxcox_trained.model_name(sarima_boxcox[0],sarima_boxcox[1])


boxcox_trained.fig()


boxcox_trained.resid_fig()


boxcox_trained.get_all()

For model ((2, 1, 0), (3, 1, 0, 7)):

Mean Squared Error (MSE):  484785323.3124968 
Root Mean Squared Error (RMSE):  22017.841022963556 
Mean Absolute Error (MAE):  17284.284351770908 
Mean Absolute Percentage Error (MAPE):  4.4323663578292125 
Mean Absolute Scaled Error (MASE):  268.55297881413225


#Preparing data into the right format to pass to scikitlearn's LinearRegression

X = [i for i in range(0, len(train_df.new_cases))]
X = np.reshape(X, (len(X), 1))


#Training a linear regression model on the series and differencing the model line from the original series

lr = LinearRegression()
lr.fit(X,train_df.new_cases)

lr_trend = lr.predict(X)
lr_series = train_df.new_cases - lr_trend
train_df['lr'] = lr_series


test_stationarity(train_df, 'lr')


adf_test(lr_series)

Results of Dickey-Fuller Test:
Test Statistic                  -2.968003
p-value                          0.037992
#Lags Used                      15.000000
Number of Observations Used    501.000000
Critical Value (1%)             -3.443470
Critical Value (5%)             -2.867326
Critical Value (10%)            -2.569852
dtype: float64


kpss_test(lr_series)

Results of KPSS Test:
Test Statistic            0.161128
p-value                   0.100000
Lags Used                14.000000
Critical Value (10%)      0.347000
Critical Value (5%)       0.463000
Critical Value (2.5%)     0.574000
Critical Value (1%)       0.739000
dtype: float64

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\stattools.py:1911: InterpolationWarning:

The test statistic is outside of the range of p-values available in the
look-up table. The actual p-value is greater than the p-value returned.


#Standard deviation of LR differenced series

train_df['lr'].std()

110541.71010978249


#Standard deviation of single order differenced LR differenced series

train_df['lrdiff1'] = train_df['lr'] - train_df['lr'].shift(1)
train_df['lrdiff1'].std()

51023.72721000706


#Standard deviation of double order differenced LR differenced series

train_df['lrdiff2'] = train_df['lrdiff1'] - train_df['lrdiff1'].shift(1)
train_df['lrdiff2'].std()

65671.92907876165


#Standard deviation of single order seasonally differenced LR differenced series

train_df['lrsdiff'] = train_df['lr'] - train_df['lr'].shift(7)
train_df['lrsdiff'].std()

50242.04381891632


#Standard deviation of single order seasonal and non-seasonal LR differenced series

train_df['lrsdiff_plus_diff'] = train_df['lrsdiff'] - train_df['lrsdiff'].shift(1)
train_df['lrsdiff_plus_diff'].std()

38450.48712209422


create_corr_plot(train_df['lrsdiff_plus_diff'])


create_corr_plot(train_df['lrsdiff_plus_diff'], plot_pacf=True)


#Training SARIMA with AR=2, I=1, S=7, S(AR)=3, S(I)=1 on LR differenced series

sarima_lr = ((2,1,0),(3,1,0,7))
with warnings.catch_warnings():
    warnings.filterwarnings("ignore")
    actual_lr, predicted_lr, model_lr = get_model_pred(sarima_lr, aic_bic=False, series_type='lr')
lr_trained = Metrics(actual_lr, predicted_lr, model_lr)
lr_trained.model_name(sarima_lr[0],sarima_lr[1])


lr_trained.fig()


lr_trained.resid_fig()


lr_trained.get_all()

For model ((2, 1, 0), (3, 1, 0, 7)):

Mean Squared Error (MSE):  552962594.59609 
Root Mean Squared Error (RMSE):  23515.156699373492 
Mean Absolute Error (MAE):  19121.640513528248 
Mean Absolute Percentage Error (MAPE):  5.019525146451417 
Mean Absolute Scaled Error (MASE):  0.5822900156199894


#Function to create grid of parameters for SARIMA grid search

p_params = list(range(6))
d_params = [0,1,2]
q_params = list(range(6))
P_params = list(range(6))
D_params = [0,1]
Q_params = list(range(6))
S_params = [0,7]

def create_param_list(p_params, d_params, q_params, P_params, D_params, Q_params, S_params, series_type='normal'):
    param_list = []
    for p in p_params:
        for d in d_params:
            for q in q_params:
                for P in P_params:
                    for D in D_params:
                        for Q in Q_params:
                            for S in S_params:
                                if D==1 and d==2:
                                    continue
                                if S==0 and (P>0 or D>0 or Q>0):
                                    continue
                                if d==0 and D==0:
                                    continue
                                param_list.append(((p,d,q),(P,D,Q,S)))
    return param_list


#Function to grid search SARIMA models across various parameters

def train_and_pred(params, aic_bic=True, split_type='val', series_type='normal'):
    p,d,q = params[0]
    P,D,Q,S = params[1]
    try:
        sarima_grid = ((p,d,q),(P,D,Q,S))
        with warnings.catch_warnings():
            warnings.filterwarnings("ignore")
            actual_grid, predicted_grid, model_grid = get_model_pred(sarima_grid, aic_bic=aic_bic,
                                                                     split_type=split_type, series_type=series_type)
        if aic_bic:
            return actual_grid, predicted_grid, model_grid
        metrics_grid = Metrics(actual_grid, predicted_grid, model_grid)
        metrics_grid.model_name((p,d,q),(P,D,Q,S))
        return metrics_grid
    except: #some models are not computable and throw errors
        return


#Function to execute the grid-search with or without parallel processing

def grid_search_sarimax(parallel=False, aic_bic=True, split_type='val', series_type='normal'):
    """
    parallel:    True uses all CPU cores.
                 False runs it normally.
    aic_bic:     True computes and returns the AIC and BIC scores for all models in param_list as class Metrics.
                 False runs a rolling forecast of 30 days on each parameter, thus training each model 30 times. It 
                 then returns a list of instances of class Metrics.
    split_type:  'val' to use train as train and first validation set as test.
                 'val2' to use (train+val) as train and second validation set as test.
                 'test' to use (train+val+val2) as train and test set as test.
    series_type: 'normal' to use the default series in SARIMA models.
                 'reg' to use series detrended by linear regression.
                 'boxcox' to use series transformed by boxcox.
    """
    param_list = create_param_list(p_params, d_params, q_params, P_params, D_params, Q_params, S_params,
                                   series_type=series_type)
    if not parallel:
        models = [train_and_pred(param, aic_bic=aic_bic, split_type=split_type, series_type=series_type) 
                  for param in tqdm(param_list)]
    else:
        from multiprocess import Pool
        import parallel #custom .py file required (in current directory) only for parallel processing
        import os
        max_pool = os.cpu_count()
        with Pool(max_pool) as p, tqdm(total=len(param_list)) as pbar:
            models = [(p.apply_async(parallel.train_and_pred, (param,),
                                     dict(aic_bic=aic_bic, split_type=split_type, series_type=series_type),
                                     callback=lambda _: pbar.update(1))) for param in param_list]
            models = [x.get() for x in models]
    return models


%%time
#Grid search on the normal series to return AIC and BIC scores

normal_aic_bic = grid_search_sarimax()

100%|███████████████████████████████████████████████████████████████████████████| 5256/5256 [16:12:38<00:00, 11.10s/it]

Wall time: 16h 12min 38s


%%time
#Grid search on the boxcox series to return AIC and BIC scores

boxcox_aic_bic = grid_search_sarimax(series_type='boxcox')

100%|███████████████████████████████████████████████████████████████████████████| 5256/5256 [21:21:27<00:00, 14.63s/it]

Wall time: 21h 21min 27s


%%time
#Grid search on the LR detrended series to return AIC and BIC scores

lr_aic_bic = grid_search_sarimax(series_type='lr')

100%|███████████████████████████████████████████████████████████████████████████| 5256/5256 [18:15:11<00:00, 12.50s/it]

Wall time: 18h 15min 11s
Parser   : 170 ms


#Function to get the best models based on IC scores in each order of differencing

def get_top_models(model_list):
    diffs_01 = []
    diffs_10 = []
    diffs_11 = []
    diffs_20 = []
    model_list = [x for x in normal_aic_bic if x!=None]
    top_model_list = []

    for model in model_list:
        if model[0][0][1] == 0:
            diffs_01.append(model)
        elif model[0][0][1] == 2:
            diffs_20.append(model)
        else:
            if model[0][1][1] == 0:
                diffs_10.append(model)
            elif model [0][1][1] == 1:
                diffs_11.append(model)

    for x in (diffs_01, diffs_10, diffs_11, diffs_20):
        temp_df = pd.DataFrame(x)
        top_model_list.extend(temp_df.sort_values(1).head(5)[0].values)
        top_model_list.extend(temp_df.sort_values(2).head(5)[0].values)
        
    return top_model_list


%%time
#The top 40 models from each series being trained to find their 30-day rolling forecast RMSE scores on 1st Validation

normal_top = []
boxcox_top = []
lr_top = []

for x in tqdm(get_top_models(normal_aic_bic)):
    normal_top.append(train_and_pred(x, aic_bic=False))

for x in tqdm(get_top_models(boxcox_aic_bic)):
    boxcox_top.append(train_and_pred(x, series_type='boxcox', aic_bic=False))
    
for x in tqdm(get_top_models(lr_aic_bic)):
    lr_top.append(train_and_pred(x, series_type='lr', aic_bic=False))

100%|███████████████████████████████████████████████████████████████████████████████| 40/40 [2:55:22<00:00, 263.07s/it]
100%|███████████████████████████████████████████████████████████████████████████████| 40/40 [2:29:22<00:00, 224.07s/it]
100%|███████████████████████████████████████████████████████████████████████████████| 40/40 [2:49:51<00:00, 254.80s/it]

Wall time: 8h 14min 40s


#Concatenating the best scores to a single dataframe

top_models_grid = pd.concat([pd.DataFrame([(x.model_attribs, x.rmse(),'normal') for x in normal_top if x!=None]),
                             pd.DataFrame([(x.model_attribs, x.rmse(),'boxcox') for x in boxcox_top if x!=None]),
                             pd.DataFrame([(x.model_attribs, x.rmse(),'lr') for x in lr_top if x!=None])])


#Manually passing the best results a variable so it can be used even if the time-taking computation is skipped

top_models_grid = pd.DataFrame([[((3, 2, 2), (2, 0, 1, 7)), 19634.74428968301, 'boxcox'],
 [((5, 1, 3), (3, 0, 2, 7)), 19745.971512098506, 'boxcox'],
 [((5, 1, 3), (2, 0, 2, 7)), 19871.168770235356, 'boxcox'],
 [((5, 1, 3), (3, 0, 1, 7)), 19895.86596117377, 'boxcox'],
 [((5, 1, 3), (1, 0, 2, 7)), 19966.745046262848, 'boxcox'],
 [((5, 1, 3), (2, 0, 1, 7)), 19993.681918591043, 'boxcox'],
 [((5, 1, 3), (2, 0, 1, 7)), 19993.681918591043, 'boxcox'],
 [((5, 1, 3), (0, 1, 3, 7)), 20164.362481277407, 'boxcox'],
 [((3, 0, 3), (2, 1, 5, 7)), 20242.323707311414, 'boxcox'],
 [((1, 0, 1), (1, 1, 1, 7)), 20270.54552704959, 'boxcox'],
 [((0, 1, 1), (1, 1, 1, 7)), 20272.70613794951, 'boxcox'],
 [((5, 1, 3), (0, 1, 2, 7)), 20292.36329395848, 'boxcox'],
 [((5, 1, 3), (3, 0, 2, 7)), 20304.77908584254, 'lr'],
 [((5, 1, 3), (1, 0, 2, 7)), 20320.4173948662, 'lr'],
 [((1, 0, 1), (0, 1, 2, 7)), 20341.326477819264, 'boxcox'],
 [((3, 0, 3), (1, 1, 1, 7)), 20345.958576452045, 'boxcox'],
 [((3, 0, 3), (1, 1, 1, 7)), 20345.958576452045, 'boxcox'],
 [((5, 1, 3), (2, 0, 1, 7)), 20357.02064064762, 'lr'],
 [((5, 1, 3), (2, 0, 1, 7)), 20357.02064064762, 'lr'],
 [((5, 1, 3), (3, 0, 2, 7)), 20366.350883787552, 'normal'],
 [((0, 1, 1), (0, 1, 2, 7)), 20395.00699393534, 'boxcox'],
 [((5, 1, 3), (2, 0, 1, 7)), 20478.591594869296, 'normal'],
 [((5, 1, 3), (2, 0, 1, 7)), 20478.591594869296, 'normal'],
 [((5, 1, 3), (2, 0, 2, 7)), 20479.8400468167, 'lr'],
 [((5, 1, 3), (1, 0, 1, 7)), 20496.169103386972, 'boxcox'],
 [((2, 1, 3), (1, 1, 1, 7)), 20498.779215374947, 'boxcox'],
 [((5, 1, 3), (1, 1, 1, 7)), 20549.802454227894, 'boxcox'],
 [((5, 1, 3), (2, 1, 1, 7)), 20551.639587278907, 'boxcox'],
 [((5, 1, 3), (1, 0, 0, 7)), 20560.296084296373, 'boxcox'],
 [((2, 0, 2), (1, 1, 1, 7)), 20574.780884399224, 'boxcox'],
 [((5, 1, 3), (2, 0, 2, 7)), 20602.21424911169, 'normal'],
 [((1, 1, 3), (1, 1, 1, 7)), 20604.58299009131, 'boxcox'],
 [((2, 0, 3), (1, 1, 1, 7)), 20608.790108674315, 'boxcox'],
 [((5, 1, 3), (1, 0, 2, 7)), 20631.662368602985, 'normal'],
 [((5, 1, 3), (3, 0, 0, 7)), 20674.344538756068, 'lr'],
 [((5, 1, 3), (3, 0, 0, 7)), 20674.344538756068, 'lr'],
 [((5, 1, 3), (1, 1, 2, 7)), 20678.90013139158, 'boxcox'],
 [((1, 1, 3), (0, 1, 2, 7)), 20688.361648607995, 'boxcox'],
 [((5, 1, 3), (1, 0, 0, 7)), 20694.940490399345, 'normal'],
 [((5, 1, 3), (3, 0, 1, 7)), 20727.08312501934, 'lr'],
 [((5, 1, 3), (1, 0, 0, 7)), 20735.185465402927, 'lr'],
 [((5, 2, 5), (1, 0, 2, 7)), 20828.976646340067, 'lr'],
 [((5, 1, 3), (3, 0, 0, 7)), 20863.152905109797, 'normal'],
 [((5, 1, 3), (3, 0, 0, 7)), 20863.152905109797, 'normal'],
 [((5, 0, 3), (1, 1, 1, 7)), 20878.732980754117, 'boxcox'],
 [((5, 2, 5), (2, 0, 1, 7)), 20902.231767221616, 'lr'],
 [((5, 1, 3), (3, 0, 1, 7)), 20951.34408129412, 'normal'],
 [((1, 2, 2), (3, 0, 1, 7)), 20962.941244034806, 'lr'],
 [((1, 2, 2), (3, 0, 1, 7)), 20962.941244034806, 'lr'],
 [((5, 1, 3), (1, 0, 1, 7)), 20978.60436714676, 'lr']])


auto_arima(train_df.new_cases, start_p=0, d=0, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           information_criterion='aic', with_intercept=False)

Performing stepwise search to minimize aic
 ARIMA(0,0,0)(0,0,0)[7]             : AIC=14847.134, Time=0.03 sec
 ARIMA(1,0,0)(1,0,0)[7]             : AIC=12417.136, Time=0.24 sec
 ARIMA(0,0,1)(0,0,1)[7]             : AIC=14398.546, Time=0.27 sec
 ARIMA(1,0,0)(0,0,0)[7]             : AIC=inf, Time=0.04 sec
 ARIMA(1,0,0)(2,0,0)[7]             : AIC=12397.918, Time=0.55 sec
 ARIMA(1,0,0)(3,0,0)[7]             : AIC=12392.508, Time=0.57 sec
 ARIMA(1,0,0)(4,0,0)[7]             : AIC=12391.667, Time=1.28 sec
 ARIMA(1,0,0)(5,0,0)[7]             : AIC=inf, Time=2.05 sec
 ARIMA(1,0,0)(4,0,1)[7]             : AIC=12386.824, Time=2.15 sec
 ARIMA(1,0,0)(3,0,1)[7]             : AIC=12384.933, Time=1.30 sec
 ARIMA(1,0,0)(2,0,1)[7]             : AIC=12383.449, Time=0.75 sec
 ARIMA(1,0,0)(1,0,1)[7]             : AIC=12385.889, Time=0.41 sec
 ARIMA(1,0,0)(2,0,2)[7]             : AIC=12384.862, Time=1.02 sec
 ARIMA(1,0,0)(1,0,2)[7]             : AIC=12384.507, Time=0.63 sec
 ARIMA(1,0,0)(3,0,2)[7]             : AIC=12386.859, Time=1.40 sec
 ARIMA(0,0,0)(2,0,1)[7]             : AIC=12632.432, Time=0.34 sec
 ARIMA(2,0,0)(2,0,1)[7]             : AIC=12351.561, Time=0.80 sec
 ARIMA(2,0,0)(1,0,1)[7]             : AIC=inf, Time=1.13 sec
 ARIMA(2,0,0)(2,0,0)[7]             : AIC=12364.041, Time=0.42 sec
 ARIMA(2,0,0)(3,0,1)[7]             : AIC=12353.467, Time=1.45 sec
 ARIMA(2,0,0)(2,0,2)[7]             : AIC=12353.433, Time=0.99 sec
 ARIMA(2,0,0)(1,0,0)[7]             : AIC=12376.542, Time=0.25 sec
 ARIMA(2,0,0)(1,0,2)[7]             : AIC=12352.533, Time=0.79 sec
 ARIMA(2,0,0)(3,0,0)[7]             : AIC=12358.682, Time=0.85 sec
 ARIMA(2,0,0)(3,0,2)[7]             : AIC=12355.367, Time=1.67 sec
 ARIMA(3,0,0)(2,0,1)[7]             : AIC=12336.969, Time=0.97 sec
 ARIMA(3,0,0)(1,0,1)[7]             : AIC=12339.843, Time=0.54 sec
 ARIMA(3,0,0)(2,0,0)[7]             : AIC=12349.523, Time=0.61 sec
 ARIMA(3,0,0)(3,0,1)[7]             : AIC=12338.811, Time=2.02 sec
 ARIMA(3,0,0)(2,0,2)[7]             : AIC=12338.757, Time=1.12 sec
 ARIMA(3,0,0)(1,0,0)[7]             : AIC=12363.305, Time=0.44 sec
 ARIMA(3,0,0)(1,0,2)[7]             : AIC=12337.979, Time=0.91 sec
 ARIMA(3,0,0)(3,0,0)[7]             : AIC=inf, Time=nan sec
 ARIMA(3,0,0)(3,0,2)[7]             : AIC=12340.682, Time=2.22 sec
 ARIMA(4,0,0)(2,0,1)[7]             : AIC=12281.747, Time=1.43 sec
 ARIMA(4,0,0)(1,0,1)[7]             : AIC=12286.243, Time=0.83 sec
 ARIMA(4,0,0)(2,0,0)[7]             : AIC=inf, Time=nan sec
 ARIMA(4,0,0)(3,0,1)[7]             : AIC=12283.657, Time=2.55 sec
 ARIMA(4,0,0)(2,0,2)[7]             : AIC=12283.617, Time=1.79 sec
 ARIMA(4,0,0)(1,0,0)[7]             : AIC=12311.497, Time=0.58 sec
 ARIMA(4,0,0)(1,0,2)[7]             : AIC=12282.934, Time=1.17 sec
 ARIMA(4,0,0)(3,0,0)[7]             : AIC=12289.908, Time=1.62 sec
 ARIMA(4,0,0)(3,0,2)[7]             : AIC=12285.520, Time=2.92 sec
 ARIMA(5,0,0)(2,0,1)[7]             : AIC=12241.872, Time=2.39 sec
 ARIMA(5,0,0)(1,0,1)[7]             : AIC=12247.523, Time=0.95 sec
 ARIMA(5,0,0)(2,0,0)[7]             : AIC=12260.272, Time=1.42 sec
 ARIMA(5,0,0)(3,0,1)[7]             : AIC=12243.756, Time=4.31 sec
 ARIMA(5,0,0)(2,0,2)[7]             : AIC=12243.709, Time=3.00 sec
 ARIMA(5,0,0)(1,0,0)[7]             : AIC=12277.175, Time=0.65 sec
 ARIMA(5,0,0)(1,0,2)[7]             : AIC=12243.308, Time=1.48 sec
 ARIMA(5,0,0)(3,0,0)[7]             : AIC=12251.491, Time=2.02 sec
 ARIMA(5,0,0)(3,0,2)[7]             : AIC=12245.810, Time=3.77 sec
 ARIMA(5,0,1)(2,0,1)[7]             : AIC=12243.800, Time=2.36 sec
 ARIMA(4,0,1)(2,0,1)[7]             : AIC=12283.770, Time=1.52 sec
 ARIMA(5,0,0)(2,0,1)[7] intercept   : AIC=12258.205, Time=2.25 sec

Best model:  ARIMA(5,0,0)(2,0,1)[7]          
Total fit time: 72.183 seconds

ARIMA(maxiter=200, order=(5, 0, 0), scoring_args={},
      seasonal_order=(2, 0, 1, 7), suppress_warnings=True,
      with_intercept=False)


auto_arima(train_df.new_cases, start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           information_criterion='aic', with_intercept=False)

Performing stepwise search to minimize aic
 ARIMA(0,1,0)(0,0,0)[7]             : AIC=12652.379, Time=0.03 sec
 ARIMA(1,1,0)(1,0,0)[7]             : AIC=12351.180, Time=0.12 sec
 ARIMA(0,1,1)(0,0,1)[7]             : AIC=12484.373, Time=0.09 sec
 ARIMA(1,1,0)(0,0,0)[7]             : AIC=12638.620, Time=0.02 sec
 ARIMA(1,1,0)(2,0,0)[7]             : AIC=12339.292, Time=0.20 sec
 ARIMA(1,1,0)(3,0,0)[7]             : AIC=12333.633, Time=0.44 sec
 ARIMA(1,1,0)(4,0,0)[7]             : AIC=inf, Time=0.75 sec
 ARIMA(1,1,0)(3,0,1)[7]             : AIC=12327.835, Time=0.79 sec
 ARIMA(1,1,0)(2,0,1)[7]             : AIC=12325.880, Time=0.40 sec
 ARIMA(1,1,0)(1,0,1)[7]             : AIC=12329.440, Time=0.15 sec
 ARIMA(1,1,0)(2,0,2)[7]             : AIC=12327.819, Time=0.59 sec
 ARIMA(1,1,0)(1,0,2)[7]             : AIC=12326.728, Time=0.34 sec
 ARIMA(1,1,0)(3,0,2)[7]             : AIC=12329.731, Time=1.08 sec
 ARIMA(0,1,0)(2,0,1)[7]             : AIC=12206.470, Time=0.27 sec
 ARIMA(0,1,0)(1,0,1)[7]             : AIC=12217.224, Time=0.13 sec
 ARIMA(0,1,0)(2,0,0)[7]             : AIC=12245.360, Time=0.10 sec
 ARIMA(0,1,0)(3,0,1)[7]             : AIC=12206.645, Time=0.61 sec
 ARIMA(0,1,0)(2,0,2)[7]             : AIC=12207.338, Time=0.56 sec
 ARIMA(0,1,0)(1,0,0)[7]             : AIC=12284.885, Time=0.04 sec
 ARIMA(0,1,0)(1,0,2)[7]             : AIC=12211.297, Time=0.25 sec
 ARIMA(0,1,0)(3,0,0)[7]             : AIC=12230.787, Time=0.25 sec
 ARIMA(0,1,0)(3,0,2)[7]             : AIC=12210.080, Time=0.62 sec
 ARIMA(0,1,1)(2,0,1)[7]             : AIC=12327.477, Time=0.36 sec
 ARIMA(1,1,1)(2,0,1)[7]             : AIC=12313.196, Time=0.57 sec
 ARIMA(0,1,0)(2,0,1)[7] intercept   : AIC=12366.108, Time=0.29 sec

Best model:  ARIMA(0,1,0)(2,0,1)[7]          
Total fit time: 9.087 seconds

ARIMA(maxiter=200, order=(0, 1, 0), scoring_args={},
      seasonal_order=(2, 0, 1, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((0,1,0),(2,0,1,7)), aic_bic=False).rmse()

22916.475623546372


auto_arima(train_df.new_cases, start_p=0, d=2, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           information_criterion='aic', with_intercept=False)

Performing stepwise search to minimize aic
 ARIMA(0,2,0)(0,0,0)[7]             : AIC=12887.709, Time=0.03 sec
 ARIMA(1,2,0)(1,0,0)[7]             : AIC=12593.491, Time=0.12 sec
 ARIMA(0,2,1)(0,0,1)[7]             : AIC=inf, Time=0.31 sec
 ARIMA(1,2,0)(0,0,0)[7]             : AIC=12846.119, Time=0.02 sec
 ARIMA(1,2,0)(2,0,0)[7]             : AIC=12582.078, Time=0.19 sec
 ARIMA(1,2,0)(3,0,0)[7]             : AIC=12575.541, Time=0.42 sec
 ARIMA(1,2,0)(4,0,0)[7]             : AIC=12574.957, Time=0.79 sec
 ARIMA(1,2,0)(5,0,0)[7]             : AIC=inf, Time=1.56 sec
 ARIMA(1,2,0)(4,0,1)[7]             : AIC=12568.056, Time=1.78 sec
 ARIMA(1,2,0)(3,0,1)[7]             : AIC=12566.179, Time=0.87 sec
 ARIMA(1,2,0)(2,0,1)[7]             : AIC=12564.179, Time=0.43 sec
 ARIMA(1,2,0)(1,0,1)[7]             : AIC=12570.765, Time=0.22 sec
 ARIMA(1,2,0)(2,0,2)[7]             : AIC=12566.179, Time=0.70 sec
 ARIMA(1,2,0)(1,0,2)[7]             : AIC=12565.003, Time=0.40 sec
 ARIMA(1,2,0)(3,0,2)[7]             : AIC=12567.640, Time=1.00 sec
 ARIMA(0,2,0)(2,0,1)[7]             : AIC=12686.634, Time=0.34 sec
 ARIMA(2,2,0)(2,0,1)[7]             : AIC=12529.595, Time=0.59 sec
 ARIMA(2,2,0)(1,0,1)[7]             : AIC=12532.720, Time=0.23 sec
 ARIMA(2,2,0)(2,0,0)[7]             : AIC=12544.028, Time=0.27 sec
 ARIMA(2,2,0)(3,0,1)[7]             : AIC=12531.488, Time=1.10 sec
 ARIMA(2,2,0)(2,0,2)[7]             : AIC=12531.471, Time=0.97 sec
 ARIMA(2,2,0)(1,0,0)[7]             : AIC=12558.755, Time=0.15 sec
 ARIMA(2,2,0)(1,0,2)[7]             : AIC=12530.472, Time=0.45 sec
 ARIMA(2,2,0)(3,0,0)[7]             : AIC=12538.803, Time=0.56 sec
 ARIMA(2,2,0)(3,0,2)[7]             : AIC=12533.499, Time=1.53 sec
 ARIMA(3,2,0)(2,0,1)[7]             : AIC=12476.774, Time=0.75 sec
 ARIMA(3,2,0)(1,0,1)[7]             : AIC=12479.897, Time=0.29 sec
 ARIMA(3,2,0)(2,0,0)[7]             : AIC=12490.905, Time=0.39 sec
 ARIMA(3,2,0)(3,0,1)[7]             : AIC=12478.752, Time=1.29 sec
 ARIMA(3,2,0)(2,0,2)[7]             : AIC=12478.743, Time=1.08 sec
 ARIMA(3,2,0)(1,0,0)[7]             : AIC=12502.605, Time=0.21 sec
 ARIMA(3,2,0)(1,0,2)[7]             : AIC=12476.979, Time=0.53 sec
 ARIMA(3,2,0)(3,0,0)[7]             : AIC=inf, Time=0.69 sec
 ARIMA(3,2,0)(3,0,2)[7]             : AIC=12480.420, Time=1.60 sec
 ARIMA(4,2,0)(2,0,1)[7]             : AIC=12444.417, Time=0.93 sec
 ARIMA(4,2,0)(1,0,1)[7]             : AIC=12447.607, Time=0.36 sec
 ARIMA(4,2,0)(2,0,0)[7]             : AIC=12458.471, Time=0.48 sec
 ARIMA(4,2,0)(3,0,1)[7]             : AIC=12446.416, Time=1.76 sec
 ARIMA(4,2,0)(2,0,2)[7]             : AIC=12446.415, Time=1.23 sec
 ARIMA(4,2,0)(1,0,0)[7]             : AIC=12471.270, Time=0.31 sec
 ARIMA(4,2,0)(1,0,2)[7]             : AIC=12444.830, Time=0.57 sec
 ARIMA(4,2,0)(3,0,0)[7]             : AIC=inf, Time=0.90 sec
 ARIMA(4,2,0)(3,0,2)[7]             : AIC=12448.204, Time=2.29 sec
 ARIMA(5,2,0)(2,0,1)[7]             : AIC=12324.513, Time=1.31 sec
 ARIMA(5,2,0)(1,0,1)[7]             : AIC=12335.549, Time=0.49 sec
 ARIMA(5,2,0)(2,0,0)[7]             : AIC=12341.989, Time=0.72 sec
 ARIMA(5,2,0)(3,0,1)[7]             : AIC=12325.688, Time=2.39 sec
 ARIMA(5,2,0)(2,0,2)[7]             : AIC=12325.478, Time=1.69 sec
 ARIMA(5,2,0)(1,0,0)[7]             : AIC=12345.520, Time=0.44 sec
 ARIMA(5,2,0)(1,0,2)[7]             : AIC=12324.496, Time=0.75 sec
 ARIMA(5,2,0)(0,0,2)[7]             : AIC=12422.931, Time=0.51 sec
 ARIMA(5,2,0)(1,0,3)[7]             : AIC=12325.773, Time=1.45 sec
 ARIMA(5,2,0)(0,0,1)[7]             : AIC=12451.384, Time=0.27 sec
 ARIMA(5,2,0)(0,0,3)[7]             : AIC=12401.735, Time=0.87 sec
 ARIMA(5,2,0)(2,0,3)[7]             : AIC=12328.496, Time=1.36 sec
 ARIMA(5,2,1)(1,0,2)[7]             : AIC=inf, Time=2.40 sec
 ARIMA(4,2,1)(1,0,2)[7]             : AIC=inf, Time=1.14 sec
 ARIMA(5,2,0)(1,0,2)[7] intercept   : AIC=12326.612, Time=0.83 sec

Best model:  ARIMA(5,2,0)(1,0,2)[7]          
Total fit time: 47.369 seconds

ARIMA(maxiter=200, order=(5, 2, 0), scoring_args={},
      seasonal_order=(1, 0, 2, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((5,2,0),(1,0,2,7)), aic_bic=False).rmse()

24093.09049382263


auto_arima(train_df.new_cases, start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=1, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           information_criterion='aic', with_intercept=False)

Performing stepwise search to minimize aic
 ARIMA(0,1,0)(0,1,0)[7]             : AIC=12192.703, Time=0.03 sec
 ARIMA(1,1,0)(1,1,0)[7]             : AIC=12029.615, Time=0.09 sec
 ARIMA(0,1,1)(0,1,1)[7]             : AIC=11973.969, Time=0.12 sec
 ARIMA(0,1,1)(0,1,0)[7]             : AIC=12062.516, Time=0.04 sec
 ARIMA(0,1,1)(1,1,1)[7]             : AIC=11959.695, Time=0.21 sec
 ARIMA(0,1,1)(1,1,0)[7]             : AIC=12010.839, Time=0.10 sec
 ARIMA(0,1,1)(2,1,1)[7]             : AIC=11961.346, Time=0.47 sec
 ARIMA(0,1,1)(1,1,2)[7]             : AIC=11961.235, Time=0.57 sec
 ARIMA(0,1,1)(0,1,2)[7]             : AIC=11962.734, Time=0.25 sec
 ARIMA(0,1,1)(2,1,0)[7]             : AIC=11987.528, Time=0.21 sec
 ARIMA(0,1,1)(2,1,2)[7]             : AIC=11963.597, Time=0.76 sec
 ARIMA(0,1,0)(1,1,1)[7]             : AIC=12042.330, Time=0.15 sec
 ARIMA(1,1,1)(1,1,1)[7]             : AIC=11960.674, Time=0.32 sec
 ARIMA(0,1,2)(1,1,1)[7]             : AIC=11963.630, Time=0.29 sec
 ARIMA(1,1,0)(1,1,1)[7]             : AIC=11973.676, Time=0.21 sec
 ARIMA(1,1,2)(1,1,1)[7]             : AIC=11960.729, Time=0.36 sec
 ARIMA(0,1,1)(1,1,1)[7] intercept   : AIC=11964.313, Time=0.22 sec

Best model:  ARIMA(0,1,1)(1,1,1)[7]          
Total fit time: 4.423 seconds

ARIMA(maxiter=200, order=(0, 1, 1), scoring_args={},
      seasonal_order=(1, 1, 1, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((0,1,1),(1,1,1,7)), aic_bic=False).rmse()

23398.170753772487


auto_arima(train_df.new_cases, start_p=0, d=0, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=1, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           information_criterion='aic', with_intercept=False)

Performing stepwise search to minimize aic
 ARIMA(0,0,0)(0,1,0)[7]             : AIC=12498.421, Time=0.04 sec
 ARIMA(1,0,0)(1,1,0)[7]             : AIC=12093.831, Time=0.14 sec
 ARIMA(0,0,1)(0,1,1)[7]             : AIC=12339.837, Time=0.10 sec
 ARIMA(1,0,0)(0,1,0)[7]             : AIC=12139.463, Time=0.02 sec
 ARIMA(1,0,0)(2,1,0)[7]             : AIC=12079.352, Time=0.32 sec
 ARIMA(1,0,0)(3,1,0)[7]             : AIC=12074.267, Time=0.67 sec
 ARIMA(1,0,0)(4,1,0)[7]             : AIC=12071.423, Time=1.36 sec
 ARIMA(1,0,0)(5,1,0)[7]             : AIC=12073.023, Time=2.01 sec
 ARIMA(1,0,0)(4,1,1)[7]             : AIC=12055.599, Time=2.40 sec
 ARIMA(1,0,0)(3,1,1)[7]             : AIC=12053.645, Time=1.02 sec
 ARIMA(1,0,0)(2,1,1)[7]             : AIC=12051.812, Time=0.48 sec
 ARIMA(1,0,0)(1,1,1)[7]             : AIC=12051.001, Time=0.23 sec
 ARIMA(1,0,0)(0,1,1)[7]             : AIC=12062.967, Time=0.17 sec
 ARIMA(1,0,0)(1,1,2)[7]             : AIC=12051.716, Time=0.61 sec
 ARIMA(1,0,0)(0,1,2)[7]             : AIC=12053.885, Time=0.41 sec
 ARIMA(1,0,0)(2,1,2)[7]             : AIC=12053.713, Time=0.86 sec
 ARIMA(0,0,0)(1,1,1)[7]             : AIC=12438.485, Time=0.13 sec
 ARIMA(2,0,0)(1,1,1)[7]             : AIC=11984.357, Time=0.29 sec
 ARIMA(2,0,0)(0,1,1)[7]             : AIC=12000.850, Time=0.22 sec
 ARIMA(2,0,0)(1,1,0)[7]             : AIC=12031.798, Time=0.18 sec
 ARIMA(2,0,0)(2,1,1)[7]             : AIC=11986.161, Time=0.71 sec
 ARIMA(2,0,0)(1,1,2)[7]             : AIC=11986.108, Time=0.69 sec
 ARIMA(2,0,0)(0,1,0)[7]             : AIC=12071.771, Time=0.03 sec
 ARIMA(2,0,0)(0,1,2)[7]             : AIC=11987.540, Time=0.39 sec
 ARIMA(2,0,0)(2,1,0)[7]             : AIC=12012.042, Time=0.41 sec
 ARIMA(2,0,0)(2,1,2)[7]             : AIC=11988.003, Time=1.18 sec
 ARIMA(3,0,0)(1,1,1)[7]             : AIC=11982.331, Time=0.43 sec
 ARIMA(3,0,0)(0,1,1)[7]             : AIC=11997.227, Time=0.28 sec
 ARIMA(3,0,0)(1,1,0)[7]             : AIC=12028.907, Time=0.26 sec
 ARIMA(3,0,0)(2,1,1)[7]             : AIC=11983.983, Time=0.80 sec
 ARIMA(3,0,0)(1,1,2)[7]             : AIC=11983.890, Time=0.82 sec
 ARIMA(3,0,0)(0,1,0)[7]             : AIC=12071.401, Time=0.05 sec
 ARIMA(3,0,0)(0,1,2)[7]             : AIC=11985.641, Time=0.49 sec
 ARIMA(3,0,0)(2,1,0)[7]             : AIC=12009.269, Time=0.52 sec
 ARIMA(3,0,0)(2,1,2)[7]             : AIC=11985.780, Time=1.26 sec
 ARIMA(4,0,0)(1,1,1)[7]             : AIC=11974.519, Time=0.57 sec
 ARIMA(4,0,0)(0,1,1)[7]             : AIC=11991.336, Time=0.32 sec
 ARIMA(4,0,0)(1,1,0)[7]             : AIC=12024.054, Time=0.34 sec
 ARIMA(4,0,0)(2,1,1)[7]             : AIC=11976.344, Time=1.06 sec
 ARIMA(4,0,0)(1,1,2)[7]             : AIC=11976.283, Time=1.02 sec
 ARIMA(4,0,0)(0,1,0)[7]             : AIC=12065.844, Time=0.05 sec
 ARIMA(4,0,0)(0,1,2)[7]             : AIC=11977.739, Time=0.63 sec
 ARIMA(4,0,0)(2,1,0)[7]             : AIC=12002.117, Time=0.78 sec
 ARIMA(4,0,0)(2,1,2)[7]             : AIC=11978.141, Time=1.98 sec
 ARIMA(5,0,0)(1,1,1)[7]             : AIC=11976.421, Time=0.72 sec
 ARIMA(4,0,1)(1,1,1)[7]             : AIC=11972.244, Time=0.77 sec
 ARIMA(4,0,1)(0,1,1)[7]             : AIC=11992.506, Time=0.59 sec
 ARIMA(4,0,1)(1,1,0)[7]             : AIC=12021.176, Time=0.74 sec
 ARIMA(4,0,1)(2,1,1)[7]             : AIC=11974.207, Time=1.87 sec
 ARIMA(4,0,1)(1,1,2)[7]             : AIC=11974.187, Time=1.79 sec
 ARIMA(4,0,1)(0,1,0)[7]             : AIC=12050.538, Time=0.41 sec
 ARIMA(4,0,1)(0,1,2)[7]             : AIC=11974.488, Time=1.26 sec
 ARIMA(4,0,1)(2,1,0)[7]             : AIC=11997.205, Time=1.47 sec
 ARIMA(4,0,1)(2,1,2)[7]             : AIC=11975.020, Time=3.36 sec
 ARIMA(3,0,1)(1,1,1)[7]             : AIC=11978.471, Time=0.80 sec
 ARIMA(5,0,1)(1,1,1)[7]             : AIC=11973.487, Time=1.28 sec
 ARIMA(4,0,2)(1,1,1)[7]             : AIC=11971.989, Time=1.32 sec
 ARIMA(4,0,2)(0,1,1)[7]             : AIC=11992.886, Time=1.09 sec
 ARIMA(4,0,2)(1,1,0)[7]             : AIC=12022.901, Time=0.93 sec
 ARIMA(4,0,2)(2,1,1)[7]             : AIC=11973.984, Time=2.57 sec
 ARIMA(4,0,2)(1,1,2)[7]             : AIC=11973.982, Time=2.10 sec
 ARIMA(4,0,2)(0,1,0)[7]             : AIC=12010.905, Time=0.96 sec
 ARIMA(4,0,2)(0,1,2)[7]             : AIC=11974.513, Time=1.78 sec
 ARIMA(4,0,2)(2,1,0)[7]             : AIC=11998.137, Time=2.04 sec
 ARIMA(4,0,2)(2,1,2)[7]             : AIC=11974.809, Time=4.23 sec
 ARIMA(3,0,2)(1,1,1)[7]             : AIC=11970.532, Time=0.91 sec
 ARIMA(3,0,2)(0,1,1)[7]             : AIC=11991.386, Time=0.70 sec
 ARIMA(3,0,2)(1,1,0)[7]             : AIC=12022.171, Time=0.67 sec
 ARIMA(3,0,2)(2,1,1)[7]             : AIC=11972.531, Time=1.63 sec
 ARIMA(3,0,2)(1,1,2)[7]             : AIC=11972.531, Time=1.67 sec
 ARIMA(3,0,2)(0,1,0)[7]             : AIC=12055.335, Time=0.57 sec
 ARIMA(3,0,2)(0,1,2)[7]             : AIC=11973.197, Time=1.09 sec
 ARIMA(3,0,2)(2,1,0)[7]             : AIC=11997.109, Time=1.76 sec
 ARIMA(3,0,2)(2,1,2)[7]             : AIC=11973.347, Time=2.64 sec
 ARIMA(2,0,2)(1,1,1)[7]             : AIC=11970.936, Time=0.73 sec
 ARIMA(3,0,3)(1,1,1)[7]             : AIC=11962.551, Time=2.03 sec
 ARIMA(3,0,3)(0,1,1)[7]             : AIC=11988.527, Time=1.30 sec
 ARIMA(3,0,3)(1,1,0)[7]             : AIC=inf, Time=2.17 sec
 ARIMA(3,0,3)(2,1,1)[7]             : AIC=11964.543, Time=2.99 sec
 ARIMA(3,0,3)(1,1,2)[7]             : AIC=11964.540, Time=4.09 sec
 ARIMA(3,0,3)(0,1,0)[7]             : AIC=inf, Time=1.26 sec
 ARIMA(3,0,3)(0,1,2)[7]             : AIC=11966.231, Time=3.23 sec
 ARIMA(3,0,3)(2,1,0)[7]             : AIC=11991.955, Time=2.83 sec
 ARIMA(3,0,3)(2,1,2)[7]             : AIC=11965.598, Time=5.22 sec
 ARIMA(2,0,3)(1,1,1)[7]             : AIC=11966.755, Time=0.97 sec
 ARIMA(4,0,3)(1,1,1)[7]             : AIC=11964.531, Time=2.70 sec
 ARIMA(3,0,4)(1,1,1)[7]             : AIC=11972.124, Time=1.87 sec
 ARIMA(2,0,4)(1,1,1)[7]             : AIC=11971.915, Time=1.06 sec
 ARIMA(4,0,4)(1,1,1)[7]             : AIC=11969.015, Time=5.16 sec
 ARIMA(3,0,3)(1,1,1)[7] intercept   : AIC=11965.197, Time=1.54 sec

Best model:  ARIMA(3,0,3)(1,1,1)[7]          
Total fit time: 109.713 seconds

ARIMA(maxiter=200, order=(3, 0, 3), scoring_args={},
      seasonal_order=(1, 1, 1, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((3,0,3),(1,1,1,7)), aic_bic=False).rmse()

23344.631391908326


auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=0, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False)

Performing stepwise search to minimize oob
 ARIMA(0,0,0)(0,0,0)[7]             : OOB=149130097925.033, Time=0.11 sec
 ARIMA(1,0,0)(1,0,0)[7]             : OOB=14519327458.053, Time=0.17 sec
 ARIMA(0,0,1)(0,0,1)[7]             : OOB=118234775301.958, Time=0.13 sec
 ARIMA(1,0,0)(0,0,0)[7]             : OOB=inf, Time=0.04 sec
 ARIMA(1,0,0)(2,0,0)[7]             : OOB=4141811267.187, Time=0.43 sec
 ARIMA(1,0,0)(3,0,0)[7]             : OOB=1672764838.984, Time=0.73 sec
 ARIMA(1,0,0)(4,0,0)[7]             : OOB=1041381702.891, Time=1.64 sec
 ARIMA(1,0,0)(5,0,0)[7]             : OOB=inf, Time=2.40 sec
 ARIMA(1,0,0)(4,0,1)[7]             : OOB=5295670313.779, Time=2.68 sec
 ARIMA(1,0,0)(3,0,1)[7]             : OOB=5870459806.856, Time=1.65 sec
 ARIMA(1,0,0)(5,0,1)[7]             : OOB=5335264072.328, Time=4.74 sec
 ARIMA(0,0,0)(4,0,0)[7]             : OOB=inf, Time=0.83 sec
 ARIMA(2,0,0)(4,0,0)[7]             : OOB=inf, Time=1.85 sec
 ARIMA(1,0,1)(4,0,0)[7]             : OOB=3563522122.458, Time=1.79 sec
 ARIMA(0,0,1)(4,0,0)[7]             : OOB=1489895420.466, Time=0.97 sec
 ARIMA(2,0,1)(4,0,0)[7]             : OOB=3937290104.746, Time=2.33 sec
 ARIMA(1,0,0)(4,0,0)[7] intercept   : OOB=4997232789.179, Time=9.50 sec

Best model:  ARIMA(1,0,0)(4,0,0)[7]          
Total fit time: 31.983 seconds

ARIMA(maxiter=200, order=(1, 0, 0), out_of_sample_size=30, scoring_args={},
      seasonal_order=(4, 0, 0, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((1,0,0),(4,0,0,7)), aic_bic=False).rmse()

22123.625625710778


auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False)

Performing stepwise search to minimize oob
 ARIMA(0,1,0)(0,0,0)[7]             : OOB=2907757359.033, Time=0.05 sec
 ARIMA(1,1,0)(1,0,0)[7]             : OOB=8627755110.970, Time=0.20 sec
 ARIMA(0,1,1)(0,0,1)[7]             : OOB=2350368698.239, Time=0.16 sec
 ARIMA(0,1,1)(0,0,0)[7]             : OOB=2900216712.111, Time=0.07 sec
 ARIMA(0,1,1)(1,0,1)[7]             : OOB=9397316159.953, Time=0.32 sec
 ARIMA(0,1,1)(0,0,2)[7]             : OOB=2185451296.647, Time=0.33 sec
 ARIMA(0,1,1)(1,0,2)[7]             : OOB=3369948122.371, Time=0.52 sec
 ARIMA(0,1,1)(0,0,3)[7]             : OOB=2652430604.514, Time=0.61 sec
 ARIMA(0,1,1)(1,0,3)[7]             : OOB=1953348658.492, Time=0.86 sec
 ARIMA(0,1,1)(2,0,3)[7]             : OOB=2067141706.594, Time=1.91 sec
 ARIMA(0,1,1)(1,0,4)[7]             : OOB=1655834724.898, Time=1.60 sec
 ARIMA(0,1,1)(0,0,4)[7]             : OOB=3332884715.906, Time=1.19 sec
 ARIMA(0,1,1)(2,0,4)[7]             : OOB=1690403002.766, Time=1.58 sec
 ARIMA(0,1,1)(1,0,5)[7]             : OOB=1730883671.167, Time=3.06 sec
 ARIMA(0,1,1)(0,0,5)[7]             : OOB=4051872458.251, Time=1.84 sec
 ARIMA(0,1,1)(2,0,5)[7]             : OOB=1667092507.939, Time=4.36 sec
 ARIMA(0,1,0)(1,0,4)[7]             : OOB=1122647948.621, Time=1.43 sec
 ARIMA(0,1,0)(0,0,4)[7]             : OOB=4624492301.100, Time=0.83 sec
 ARIMA(0,1,0)(1,0,3)[7]             : OOB=1262062747.826, Time=1.02 sec
 ARIMA(0,1,0)(2,0,4)[7]             : OOB=1200160667.520, Time=2.17 sec
 ARIMA(0,1,0)(1,0,5)[7]             : OOB=1063867969.612, Time=3.07 sec
 ARIMA(0,1,0)(0,0,5)[7]             : OOB=4792763473.571, Time=1.54 sec
 ARIMA(0,1,0)(2,0,5)[7]             : OOB=1169810123.516, Time=3.88 sec
 ARIMA(1,1,0)(1,0,5)[7]             : OOB=1472614880.313, Time=3.15 sec
 ARIMA(1,1,1)(1,0,5)[7]             : OOB=1544032217.664, Time=3.97 sec
 ARIMA(0,1,0)(1,0,5)[7] intercept   : OOB=1973576400.879, Time=2.98 sec

Best model:  ARIMA(0,1,0)(1,0,5)[7]          
Total fit time: 42.730 seconds

ARIMA(maxiter=200, order=(0, 1, 0), out_of_sample_size=30, scoring_args={},
      seasonal_order=(1, 0, 5, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((0,1,0),(1,0,5,7)), aic_bic=False).rmse()

22969.0275034592


auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=2, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False)

Performing stepwise search to minimize oob
 ARIMA(0,2,0)(0,0,0)[7]             : OOB=49341759507.333, Time=0.06 sec
 ARIMA(1,2,0)(1,0,0)[7]             : OOB=33812091417.894, Time=0.19 sec
 ARIMA(0,2,1)(0,0,1)[7]             : OOB=inf, Time=0.38 sec
 ARIMA(1,2,0)(0,0,0)[7]             : OOB=277083617242.699, Time=0.05 sec
 ARIMA(1,2,0)(2,0,0)[7]             : OOB=48143686578.797, Time=0.33 sec
 ARIMA(1,2,0)(1,0,1)[7]             : OOB=4936981762.797, Time=0.34 sec
 ARIMA(1,2,0)(0,0,1)[7]             : OOB=117493126463.159, Time=0.17 sec
 ARIMA(1,2,0)(2,0,1)[7]             : OOB=2025676562.876, Time=0.61 sec
 ARIMA(1,2,0)(3,0,1)[7]             : OOB=1991259038.332, Time=1.25 sec
 ARIMA(1,2,0)(3,0,0)[7]             : OOB=1633138827.354, Time=0.64 sec
 ARIMA(1,2,0)(4,0,0)[7]             : OOB=3446378487.520, Time=1.29 sec
 ARIMA(1,2,0)(4,0,1)[7]             : OOB=2598007577.790, Time=2.20 sec
 ARIMA(0,2,0)(3,0,0)[7]             : OOB=8775762823.245, Time=0.57 sec
 ARIMA(2,2,0)(3,0,0)[7]             : OOB=16594354765.456, Time=0.97 sec
 ARIMA(1,2,1)(3,0,0)[7]             : OOB=inf, Time=1.19 sec
 ARIMA(0,2,1)(3,0,0)[7]             : OOB=inf, Time=0.96 sec
 ARIMA(2,2,1)(3,0,0)[7]             : OOB=inf, Time=1.51 sec
 ARIMA(1,2,0)(3,0,0)[7] intercept   : OOB=2847545205.372, Time=0.72 sec

Best model:  ARIMA(1,2,0)(3,0,0)[7]          
Total fit time: 13.430 seconds

ARIMA(maxiter=200, order=(1, 2, 0), out_of_sample_size=30, scoring_args={},
      seasonal_order=(3, 0, 0, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((1,2,0),(3,0,0,7)), aic_bic=False).rmse()

29687.751741852746


auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=1, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False)

Performing stepwise search to minimize oob
 ARIMA(0,1,0)(0,1,0)[7]             : OOB=15416368714.367, Time=0.06 sec
 ARIMA(1,1,0)(1,1,0)[7]             : OOB=17276749347.203, Time=0.18 sec
 ARIMA(0,1,1)(0,1,1)[7]             : OOB=10456550693.569, Time=0.24 sec
 ARIMA(0,1,1)(0,1,0)[7]             : OOB=22555891059.215, Time=0.09 sec
 ARIMA(0,1,1)(1,1,1)[7]             : OOB=1558180381.616, Time=0.37 sec
 ARIMA(0,1,1)(1,1,0)[7]             : OOB=20295837051.377, Time=0.24 sec
 ARIMA(0,1,1)(2,1,1)[7]             : OOB=1556228839.414, Time=0.79 sec
 ARIMA(0,1,1)(2,1,0)[7]             : OOB=21629859597.917, Time=0.42 sec
 ARIMA(0,1,1)(3,1,1)[7]             : OOB=1842119099.850, Time=1.51 sec
 ARIMA(0,1,1)(2,1,2)[7]             : OOB=1568382379.992, Time=1.10 sec
 ARIMA(0,1,1)(1,1,2)[7]             : OOB=1604827627.095, Time=0.92 sec
 ARIMA(0,1,1)(3,1,0)[7]             : OOB=18382330909.786, Time=0.81 sec
 ARIMA(0,1,1)(3,1,2)[7]             : OOB=1636033463.728, Time=1.87 sec
 ARIMA(0,1,0)(2,1,1)[7]             : OOB=1655346689.754, Time=0.68 sec
 ARIMA(1,1,1)(2,1,1)[7]             : OOB=1521981850.379, Time=1.02 sec
 ARIMA(1,1,1)(1,1,1)[7]             : OOB=1530781725.977, Time=0.52 sec
 ARIMA(1,1,1)(2,1,0)[7]             : OOB=21760592751.809, Time=0.64 sec
 ARIMA(1,1,1)(3,1,1)[7]             : OOB=1803432287.608, Time=1.75 sec
 ARIMA(1,1,1)(2,1,2)[7]             : OOB=1539902117.897, Time=1.27 sec
 ARIMA(1,1,1)(1,1,0)[7]             : OOB=20454554011.088, Time=0.34 sec
 ARIMA(1,1,1)(1,1,2)[7]             : OOB=1577916328.480, Time=1.18 sec
 ARIMA(1,1,1)(3,1,0)[7]             : OOB=18434793119.136, Time=1.19 sec
 ARIMA(1,1,1)(3,1,2)[7]             : OOB=1598885035.005, Time=2.20 sec
 ARIMA(1,1,0)(2,1,1)[7]             : OOB=1515118308.985, Time=0.81 sec
 ARIMA(1,1,0)(1,1,1)[7]             : OOB=1537705393.372, Time=0.37 sec
 ARIMA(1,1,0)(2,1,0)[7]             : OOB=17489815051.118, Time=0.42 sec
 ARIMA(1,1,0)(3,1,1)[7]             : OOB=1588633611.045, Time=1.50 sec
 ARIMA(1,1,0)(2,1,2)[7]             : OOB=1546050781.414, Time=1.14 sec
 ARIMA(1,1,0)(1,1,2)[7]             : OOB=1512583823.170, Time=0.88 sec
 ARIMA(1,1,0)(0,1,2)[7]             : OOB=1952771954.390, Time=0.52 sec
 ARIMA(1,1,0)(1,1,3)[7]             : OOB=1588849446.861, Time=1.71 sec
 ARIMA(1,1,0)(0,1,1)[7]             : OOB=8140180059.525, Time=0.23 sec
 ARIMA(1,1,0)(0,1,3)[7]             : OOB=1649783921.245, Time=0.94 sec
 ARIMA(1,1,0)(2,1,3)[7]             : OOB=1539692473.681, Time=2.42 sec
 ARIMA(0,1,0)(1,1,2)[7]             : OOB=1593889249.736, Time=1.00 sec
 ARIMA(2,1,0)(1,1,2)[7]             : OOB=1574424425.355, Time=1.12 sec
 ARIMA(2,1,1)(1,1,2)[7]             : OOB=1524452938.447, Time=1.81 sec
 ARIMA(1,1,0)(1,1,2)[7] intercept   : OOB=2913045826.909, Time=1.75 sec

Best model:  ARIMA(1,1,0)(1,1,2)[7]          
Total fit time: 36.051 seconds

ARIMA(maxiter=200, order=(1, 1, 0), out_of_sample_size=30, scoring_args={},
      seasonal_order=(1, 1, 2, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((1,1,0),(1,1,2,7)), aic_bic=False).rmse()

24018.76550528243


auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=0, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=1, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False)

Performing stepwise search to minimize oob
 ARIMA(0,0,0)(0,1,0)[7]             : OOB=13300335974.367, Time=0.04 sec
 ARIMA(1,0,0)(1,1,0)[7]             : OOB=7657827237.787, Time=0.29 sec
 ARIMA(0,0,1)(0,1,1)[7]             : OOB=11073661147.179, Time=0.23 sec
 ARIMA(1,0,0)(0,1,0)[7]             : OOB=7395115274.975, Time=0.07 sec
 ARIMA(1,0,0)(0,1,1)[7]             : OOB=15315783301.960, Time=0.32 sec
 ARIMA(1,0,0)(1,1,1)[7]             : OOB=20744909212.499, Time=0.61 sec
 ARIMA(2,0,0)(0,1,0)[7]             : OOB=2542411081.362, Time=0.13 sec
 ARIMA(2,0,0)(1,1,0)[7]             : OOB=1966138188.584, Time=0.39 sec
 ARIMA(2,0,0)(2,1,0)[7]             : OOB=1614030421.149, Time=0.71 sec
 ARIMA(2,0,0)(3,1,0)[7]             : OOB=2098157617.800, Time=1.49 sec
 ARIMA(2,0,0)(2,1,1)[7]             : OOB=8196782991.069, Time=1.06 sec
 ARIMA(2,0,0)(1,1,1)[7]             : OOB=8619564853.416, Time=0.46 sec
 ARIMA(2,0,0)(3,1,1)[7]             : OOB=7138211662.535, Time=1.65 sec
 ARIMA(1,0,0)(2,1,0)[7]             : OOB=8505563678.316, Time=0.56 sec
 ARIMA(3,0,0)(2,1,0)[7]             : OOB=908479238.168, Time=0.82 sec
 ARIMA(3,0,0)(1,1,0)[7]             : OOB=1061052272.677, Time=0.56 sec
 ARIMA(3,0,0)(3,1,0)[7]             : OOB=1195484045.243, Time=1.89 sec
 ARIMA(3,0,0)(2,1,1)[7]             : OOB=5945213455.169, Time=1.25 sec
 ARIMA(3,0,0)(1,1,1)[7]             : OOB=6427321330.287, Time=0.64 sec
 ARIMA(3,0,0)(3,1,1)[7]             : OOB=5037462141.330, Time=2.17 sec
 ARIMA(4,0,0)(2,1,0)[7]             : OOB=1574860450.516, Time=1.36 sec
 ARIMA(3,0,1)(2,1,0)[7]             : OOB=2045141294.167, Time=2.30 sec
 ARIMA(2,0,1)(2,1,0)[7]             : OOB=1740925962.621, Time=1.44 sec
 ARIMA(4,0,1)(2,1,0)[7]             : OOB=889935991.426, Time=2.82 sec
 ARIMA(4,0,1)(1,1,0)[7]             : OOB=752886631.107, Time=1.43 sec
 ARIMA(4,0,1)(0,1,0)[7]             : OOB=890103611.509, Time=0.58 sec
 ARIMA(4,0,1)(1,1,1)[7]             : OOB=4772412157.911, Time=1.05 sec
 ARIMA(4,0,1)(0,1,1)[7]             : OOB=1164008782.749, Time=0.95 sec
 ARIMA(4,0,1)(2,1,1)[7]             : OOB=4936835409.304, Time=2.65 sec
 ARIMA(3,0,1)(1,1,0)[7]             : OOB=933589634.900, Time=0.88 sec
 ARIMA(4,0,0)(1,1,0)[7]             : OOB=801107692.312, Time=0.70 sec
 ARIMA(5,0,1)(1,1,0)[7]             : OOB=734061203.974, Time=1.59 sec
 ARIMA(5,0,1)(0,1,0)[7]             : OOB=1108204383.120, Time=0.72 sec
 ARIMA(5,0,1)(2,1,0)[7]             : OOB=1066732097.373, Time=2.38 sec
 ARIMA(5,0,1)(1,1,1)[7]             : OOB=4112744276.796, Time=1.59 sec
 ARIMA(5,0,1)(0,1,1)[7]             : OOB=1162887265.154, Time=1.06 sec
 ARIMA(5,0,1)(2,1,1)[7]             : OOB=4222679035.768, Time=3.36 sec
 ARIMA(5,0,0)(1,1,0)[7]             : OOB=753682263.156, Time=0.71 sec
 ARIMA(5,0,2)(1,1,0)[7]             : OOB=807312101.239, Time=1.55 sec
 ARIMA(4,0,2)(1,1,0)[7]             : OOB=754564002.277, Time=1.53 sec
 ARIMA(5,0,1)(1,1,0)[7] intercept   : OOB=797351306.811, Time=1.42 sec

Best model:  ARIMA(5,0,1)(1,1,0)[7]          
Total fit time: 47.422 seconds

ARIMA(maxiter=200, order=(5, 0, 1), out_of_sample_size=30, scoring_args={},
      seasonal_order=(1, 1, 0, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((5,0,1),(1,1,0,7)), aic_bic=False).rmse()

22576.077612634832


#Might take some time to finish

auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=0, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False, stepwise=False, max_order=8)

 ARIMA(0,0,0)(0,0,0)[7]             : OOB=149130097925.033, Time=0.06 sec
 ARIMA(0,0,0)(0,0,1)[7]             : OOB=122969618541.939, Time=0.12 sec
 ARIMA(0,0,0)(0,0,2)[7]             : OOB=98911042813.291, Time=0.23 sec
 ARIMA(0,0,0)(0,0,3)[7]             : OOB=81010814631.712, Time=0.42 sec
 ARIMA(0,0,0)(0,0,4)[7]             : OOB=65503111509.395, Time=0.88 sec
 ARIMA(0,0,0)(0,0,5)[7]             : OOB=50446412859.777, Time=1.56 sec
 ARIMA(0,0,0)(1,0,0)[7]             : OOB=inf, Time=0.12 sec
 ARIMA(0,0,0)(1,0,1)[7]             : OOB=6274373572.616, Time=0.20 sec
 ARIMA(0,0,0)(1,0,2)[7]             : OOB=4392134610.864, Time=0.33 sec
 ARIMA(0,0,0)(1,0,3)[7]             : OOB=1887660056.984, Time=0.72 sec
 ARIMA(0,0,0)(1,0,4)[7]             : OOB=1002535087.430, Time=1.23 sec
 ARIMA(0,0,0)(1,0,5)[7]             : OOB=1403155293.097, Time=4.24 sec
 ARIMA(0,0,0)(2,0,0)[7]             : OOB=inf, Time=0.26 sec
 ARIMA(0,0,0)(2,0,1)[7]             : OOB=1270462254.304, Time=0.59 sec
 ARIMA(0,0,0)(2,0,2)[7]             : OOB=1233561176.390, Time=0.47 sec
 ARIMA(0,0,0)(2,0,3)[7]             : OOB=1422319980.322, Time=1.08 sec
 ARIMA(0,0,0)(2,0,4)[7]             : OOB=1343877196.191, Time=2.38 sec
 ARIMA(0,0,0)(2,0,5)[7]             : OOB=1696738195.866, Time=6.21 sec
 ARIMA(0,0,0)(3,0,0)[7]             : OOB=inf, Time=0.67 sec
 ARIMA(0,0,0)(3,0,1)[7]             : OOB=1250899228.560, Time=0.86 sec
 ARIMA(0,0,0)(3,0,2)[7]             : OOB=2622520120.229, Time=2.21 sec
 ARIMA(0,0,0)(3,0,3)[7]             : OOB=2403522608.683, Time=2.60 sec
 ARIMA(0,0,0)(3,0,4)[7]             : OOB=2692968874.469, Time=5.35 sec
 ARIMA(0,0,0)(3,0,5)[7]             : OOB=1951032344.243, Time=10.22 sec
 ARIMA(0,0,0)(4,0,0)[7]             : OOB=inf, Time=1.30 sec
 ARIMA(0,0,0)(4,0,1)[7]             : OOB=1364409860.450, Time=1.48 sec
 ARIMA(0,0,0)(4,0,2)[7]             : OOB=2486939559.152, Time=4.34 sec
 ARIMA(0,0,0)(4,0,3)[7]             : OOB=2644090243.822, Time=4.94 sec
 ARIMA(0,0,0)(4,0,4)[7]             : OOB=2729894940.256, Time=7.11 sec
 ARIMA(0,0,0)(5,0,0)[7]             : OOB=inf, Time=1.13 sec
 ARIMA(0,0,0)(5,0,1)[7]             : OOB=818960652.840, Time=5.05 sec
 ARIMA(0,0,0)(5,0,2)[7]             : OOB=2751274392.670, Time=6.87 sec
 ARIMA(0,0,0)(5,0,3)[7]             : OOB=2783966036.685, Time=12.89 sec
 ARIMA(0,0,1)(0,0,0)[7]             : OOB=143954807669.174, Time=0.11 sec
 ARIMA(0,0,1)(0,0,1)[7]             : OOB=118234775301.958, Time=0.22 sec
 ARIMA(0,0,1)(0,0,2)[7]             : OOB=96049721640.290, Time=0.42 sec
 ARIMA(0,0,1)(0,0,3)[7]             : OOB=78964041341.390, Time=0.90 sec
 ARIMA(0,0,1)(0,0,4)[7]             : OOB=64037839145.459, Time=1.44 sec
 ARIMA(0,0,1)(0,0,5)[7]             : OOB=50251118201.646, Time=3.39 sec
 ARIMA(0,0,1)(1,0,0)[7]             : OOB=3281188949.659, Time=0.18 sec
 ARIMA(0,0,1)(1,0,1)[7]             : OOB=1464445773.364, Time=0.24 sec
 ARIMA(0,0,1)(1,0,2)[7]             : OOB=1544978649.393, Time=0.50 sec
 ARIMA(0,0,1)(1,0,3)[7]             : OOB=1970259355.706, Time=1.04 sec
 ARIMA(0,0,1)(1,0,4)[7]             : OOB=2409508493.793, Time=1.95 sec
 ARIMA(0,0,1)(1,0,5)[7]             : OOB=2367345249.083, Time=4.67 sec
 ARIMA(0,0,1)(2,0,0)[7]             : OOB=1736279509.611, Time=0.59 sec
 ARIMA(0,0,1)(2,0,1)[7]             : OOB=1747287560.997, Time=0.91 sec
 ARIMA(0,0,1)(2,0,2)[7]             : OOB=1787992288.856, Time=0.95 sec
 ARIMA(0,0,1)(2,0,3)[7]             : OOB=1705764537.098, Time=1.94 sec
 ARIMA(0,0,1)(2,0,4)[7]             : OOB=1697508709.349, Time=2.73 sec
 ARIMA(0,0,1)(2,0,5)[7]             : OOB=1698903795.623, Time=4.22 sec
 ARIMA(0,0,1)(3,0,0)[7]             : OOB=1486687931.875, Time=1.01 sec
 ARIMA(0,0,1)(3,0,1)[7]             : OOB=1750981799.617, Time=1.48 sec
 ARIMA(0,0,1)(3,0,2)[7]             : OOB=1762087716.041, Time=1.49 sec
 ARIMA(0,0,1)(3,0,3)[7]             : OOB=1223173490.030, Time=4.14 sec
 ARIMA(0,0,1)(3,0,4)[7]             : OOB=1213895386.670, Time=6.50 sec
 ARIMA(0,0,1)(4,0,0)[7]             : OOB=1489895420.466, Time=1.32 sec
 ARIMA(0,0,1)(4,0,1)[7]             : OOB=1688211198.680, Time=1.62 sec
 ARIMA(0,0,1)(4,0,2)[7]             : OOB=1701985520.422, Time=2.77 sec
 ARIMA(0,0,1)(4,0,3)[7]             : OOB=1210973200.483, Time=7.33 sec
 ARIMA(0,0,1)(5,0,0)[7]             : OOB=1536067217.548, Time=2.59 sec
 ARIMA(0,0,1)(5,0,1)[7]             : OOB=1614820722.677, Time=5.74 sec
 ARIMA(0,0,1)(5,0,2)[7]             : OOB=1218631517.784, Time=10.35 sec
 ARIMA(0,0,2)(0,0,0)[7]             : OOB=139714543994.912, Time=0.17 sec
 ARIMA(0,0,2)(0,0,1)[7]             : OOB=115830203453.694, Time=0.49 sec
 ARIMA(0,0,2)(0,0,2)[7]             : OOB=97020816880.061, Time=0.92 sec
 ARIMA(0,0,2)(0,0,3)[7]             : OOB=82048837318.468, Time=1.47 sec
 ARIMA(0,0,2)(0,0,4)[7]             : OOB=67477684529.700, Time=2.89 sec
 ARIMA(0,0,2)(0,0,5)[7]             : OOB=54903431523.930, Time=4.76 sec
 ARIMA(0,0,2)(1,0,0)[7]             : OOB=8924596446.882, Time=1.02 sec
 ARIMA(0,0,2)(1,0,1)[7]             : OOB=1820561270.414, Time=0.48 sec
 ARIMA(0,0,2)(1,0,2)[7]             : OOB=1417127591.009, Time=0.91 sec
 ARIMA(0,0,2)(1,0,3)[7]             : OOB=1519620057.268, Time=1.75 sec
 ARIMA(0,0,2)(1,0,4)[7]             : OOB=1432681172.199, Time=3.26 sec
 ARIMA(0,0,2)(1,0,5)[7]             : OOB=1175612558.583, Time=6.29 sec
 ARIMA(0,0,2)(2,0,0)[7]             : OOB=5165339077.474, Time=0.95 sec
 ARIMA(0,0,2)(2,0,1)[7]             : OOB=1182877128.960, Time=1.63 sec
 ARIMA(0,0,2)(2,0,2)[7]             : OOB=1041102919.850, Time=1.64 sec
 ARIMA(0,0,2)(2,0,3)[7]             : OOB=1041014767.904, Time=2.78 sec
 ARIMA(0,0,2)(2,0,4)[7]             : OOB=1174533375.020, Time=3.33 sec
 ARIMA(0,0,2)(3,0,0)[7]             : OOB=3049874196.494, Time=1.40 sec
 ARIMA(0,0,2)(3,0,1)[7]             : OOB=973141264.573, Time=1.88 sec
 ARIMA(0,0,2)(3,0,2)[7]             : OOB=1145064701.673, Time=2.00 sec
 ARIMA(0,0,2)(3,0,3)[7]             : OOB=961993857.069, Time=3.32 sec
 ARIMA(0,0,2)(4,0,0)[7]             : OOB=2178649593.135, Time=2.79 sec
 ARIMA(0,0,2)(4,0,1)[7]             : OOB=1025039313.423, Time=3.32 sec
 ARIMA(0,0,2)(4,0,2)[7]             : OOB=1078434376.490, Time=3.18 sec
 ARIMA(0,0,2)(5,0,0)[7]             : OOB=1671461563.920, Time=3.30 sec
 ARIMA(0,0,2)(5,0,1)[7]             : OOB=1024133803.506, Time=8.69 sec
 ARIMA(0,0,3)(0,0,0)[7]             : OOB=133850476024.481, Time=0.25 sec
 ARIMA(0,0,3)(0,0,1)[7]             : OOB=113475005056.655, Time=0.54 sec
 ARIMA(0,0,3)(0,0,2)[7]             : OOB=96629260828.555, Time=0.85 sec
 ARIMA(0,0,3)(0,0,3)[7]             : OOB=82534714473.826, Time=1.75 sec
 ARIMA(0,0,3)(0,0,4)[7]             : OOB=69834057800.519, Time=3.65 sec
 ARIMA(0,0,3)(0,0,5)[7]             : OOB=59167575614.611, Time=4.26 sec
 ARIMA(0,0,3)(1,0,0)[7]             : OOB=23830743579.275, Time=0.50 sec
 ARIMA(0,0,3)(1,0,1)[7]             : OOB=5395304645.286, Time=0.45 sec
 ARIMA(0,0,3)(1,0,2)[7]             : OOB=1947189077.903, Time=0.74 sec
 ARIMA(0,0,3)(1,0,3)[7]             : OOB=1072366551.927, Time=1.48 sec
 ARIMA(0,0,3)(1,0,4)[7]             : OOB=888305239.535, Time=2.99 sec
 ARIMA(0,0,3)(2,0,0)[7]             : OOB=14304201008.108, Time=1.35 sec
 ARIMA(0,0,3)(2,0,1)[7]             : OOB=1026090494.624, Time=0.88 sec
 ARIMA(0,0,3)(2,0,2)[7]             : OOB=1260190778.527, Time=1.08 sec
 ARIMA(0,0,3)(2,0,3)[7]             : OOB=1269323393.278, Time=2.28 sec
 ARIMA(0,0,3)(3,0,0)[7]             : OOB=9249271311.502, Time=1.77 sec
 ARIMA(0,0,3)(3,0,1)[7]             : OOB=1220842308.153, Time=1.45 sec
 ARIMA(0,0,3)(3,0,2)[7]             : OOB=1189625284.435, Time=1.56 sec
 ARIMA(0,0,3)(4,0,0)[7]             : OOB=6396302410.242, Time=2.61 sec
 ARIMA(0,0,3)(4,0,1)[7]             : OOB=1271406851.995, Time=2.09 sec
 ARIMA(0,0,3)(5,0,0)[7]             : OOB=4508626210.113, Time=3.79 sec
 ARIMA(0,0,4)(0,0,0)[7]             : OOB=133610446191.952, Time=0.22 sec
 ARIMA(0,0,4)(0,0,1)[7]             : OOB=113864781727.829, Time=0.34 sec
 ARIMA(0,0,4)(0,0,2)[7]             : OOB=99062725710.128, Time=0.68 sec
 ARIMA(0,0,4)(0,0,3)[7]             : OOB=85797893589.855, Time=1.39 sec
 ARIMA(0,0,4)(0,0,4)[7]             : OOB=72743024543.089, Time=2.79 sec
 ARIMA(0,0,4)(1,0,0)[7]             : OOB=41340127200.697, Time=0.40 sec
 ARIMA(0,0,4)(1,0,1)[7]             : OOB=4681188137.976, Time=0.55 sec
 ARIMA(0,0,4)(1,0,2)[7]             : OOB=1967136410.682, Time=0.96 sec
 ARIMA(0,0,4)(1,0,3)[7]             : OOB=1364351055.079, Time=1.79 sec
 ARIMA(0,0,4)(2,0,0)[7]             : OOB=inf, Time=2.23 sec
 ARIMA(0,0,4)(2,0,1)[7]             : OOB=1413368132.239, Time=1.25 sec
 ARIMA(0,0,4)(2,0,2)[7]             : OOB=1847400362.224, Time=1.96 sec
 ARIMA(0,0,4)(3,0,0)[7]             : OOB=13925600951.330, Time=6.13 sec
 ARIMA(0,0,4)(3,0,1)[7]             : OOB=1704663105.496, Time=1.73 sec
 ARIMA(0,0,4)(4,0,0)[7]             : OOB=9604171604.555, Time=9.24 sec
 ARIMA(0,0,5)(0,0,0)[7]             : OOB=127859024967.069, Time=0.37 sec
 ARIMA(0,0,5)(0,0,1)[7]             : OOB=114268016959.399, Time=0.68 sec
 ARIMA(0,0,5)(0,0,2)[7]             : OOB=97003003121.605, Time=0.93 sec
 ARIMA(0,0,5)(0,0,3)[7]             : OOB=83396929978.197, Time=1.55 sec
 ARIMA(0,0,5)(1,0,0)[7]             : OOB=12623217238.240, Time=0.74 sec
 ARIMA(0,0,5)(1,0,1)[7]             : OOB=3152793202.502, Time=0.55 sec
 ARIMA(0,0,5)(1,0,2)[7]             : OOB=1889449046.254, Time=1.22 sec
 ARIMA(0,0,5)(2,0,0)[7]             : OOB=11313419577.337, Time=4.48 sec
 ARIMA(0,0,5)(2,0,1)[7]             : OOB=1397136067.307, Time=1.60 sec
 ARIMA(0,0,5)(3,0,0)[7]             : OOB=15357434814.680, Time=8.34 sec
 ARIMA(1,0,0)(0,0,0)[7]             : OOB=inf, Time=0.04 sec
 ARIMA(1,0,0)(0,0,1)[7]             : OOB=8161251324.884, Time=0.46 sec
 ARIMA(1,0,0)(0,0,2)[7]             : OOB=11337885583.361, Time=0.52 sec
 ARIMA(1,0,0)(0,0,3)[7]             : OOB=13962897601.467, Time=1.10 sec
 ARIMA(1,0,0)(0,0,4)[7]             : OOB=15675167359.488, Time=1.96 sec
 ARIMA(1,0,0)(0,0,5)[7]             : OOB=16085976100.969, Time=5.54 sec
 ARIMA(1,0,0)(1,0,0)[7]             : OOB=14519327458.053, Time=0.22 sec
 ARIMA(1,0,0)(1,0,1)[7]             : OOB=1531764823.286, Time=0.62 sec
 ARIMA(1,0,0)(1,0,2)[7]             : OOB=6041983272.472, Time=1.34 sec
 ARIMA(1,0,0)(1,0,3)[7]             : OOB=6795739855.641, Time=2.92 sec
 ARIMA(1,0,0)(1,0,4)[7]             : OOB=6209353505.634, Time=2.86 sec
 ARIMA(1,0,0)(1,0,5)[7]             : OOB=5510575902.925, Time=6.86 sec
 ARIMA(1,0,0)(2,0,0)[7]             : OOB=4141811267.187, Time=0.51 sec
 ARIMA(1,0,0)(2,0,1)[7]             : OOB=6736251054.478, Time=0.92 sec
 ARIMA(1,0,0)(2,0,2)[7]             : OOB=5438103038.353, Time=1.15 sec
 ARIMA(1,0,0)(2,0,3)[7]             : OOB=5428542646.247, Time=2.14 sec
 ARIMA(1,0,0)(2,0,4)[7]             : OOB=5248798786.981, Time=4.24 sec
 ARIMA(1,0,0)(2,0,5)[7]             : OOB=4992669750.097, Time=9.87 sec
 ARIMA(1,0,0)(3,0,0)[7]             : OOB=1672764838.984, Time=1.21 sec
 ARIMA(1,0,0)(3,0,1)[7]             : OOB=5870459806.856, Time=1.75 sec
 ARIMA(1,0,0)(3,0,2)[7]             : OOB=5435119106.517, Time=1.94 sec
 ARIMA(1,0,0)(3,0,3)[7]             : OOB=4149594155.024, Time=3.34 sec
 ARIMA(1,0,0)(3,0,4)[7]             : OOB=3766048124.137, Time=5.70 sec
 ARIMA(1,0,0)(4,0,0)[7]             : OOB=1041381702.891, Time=1.61 sec
 ARIMA(1,0,0)(4,0,1)[7]             : OOB=5295670313.779, Time=2.78 sec
 ARIMA(1,0,0)(4,0,2)[7]             : OOB=5624144589.522, Time=4.18 sec
 ARIMA(1,0,0)(4,0,3)[7]             : OOB=3856496184.140, Time=6.75 sec
 ARIMA(1,0,0)(5,0,0)[7]             : OOB=inf, Time=2.49 sec
 ARIMA(1,0,0)(5,0,1)[7]             : OOB=5335264072.328, Time=5.11 sec
 ARIMA(1,0,0)(5,0,2)[7]             : OOB=4614767381.909, Time=9.43 sec
 ARIMA(1,0,1)(0,0,0)[7]             : OOB=6520193305.479, Time=0.10 sec
 ARIMA(1,0,1)(0,0,1)[7]             : OOB=6920869772.064, Time=0.47 sec
 ARIMA(1,0,1)(0,0,2)[7]             : OOB=6586484417.607, Time=0.64 sec
 ARIMA(1,0,1)(0,0,3)[7]             : OOB=6623239432.208, Time=0.80 sec
 ARIMA(1,0,1)(0,0,4)[7]             : OOB=7626545628.026, Time=2.69 sec
 ARIMA(1,0,1)(0,0,5)[7]             : OOB=8520680410.621, Time=3.14 sec
 ARIMA(1,0,1)(1,0,0)[7]             : OOB=9310812106.492, Time=0.24 sec
 ARIMA(1,0,1)(1,0,1)[7]             : OOB=2255528332.175, Time=1.39 sec
 ARIMA(1,0,1)(1,0,2)[7]             : OOB=987579444.122, Time=0.71 sec
 ARIMA(1,0,1)(1,0,3)[7]             : OOB=1456607859.358, Time=1.73 sec
 ARIMA(1,0,1)(1,0,4)[7]             : OOB=1383788064.525, Time=2.85 sec
 ARIMA(1,0,1)(1,0,5)[7]             : OOB=1155844575.408, Time=5.35 sec
 ARIMA(1,0,1)(2,0,0)[7]             : OOB=6408870718.380, Time=0.68 sec
 ARIMA(1,0,1)(2,0,1)[7]             : OOB=1453893215.482, Time=0.73 sec
 ARIMA(1,0,1)(2,0,2)[7]             : OOB=1168998138.880, Time=1.19 sec
 ARIMA(1,0,1)(2,0,3)[7]             : OOB=1007280836.619, Time=1.79 sec
 ARIMA(1,0,1)(2,0,4)[7]             : OOB=920189034.497, Time=3.63 sec
 ARIMA(1,0,1)(3,0,0)[7]             : OOB=5046867362.235, Time=0.95 sec
 ARIMA(1,0,1)(3,0,1)[7]             : OOB=1299666875.633, Time=1.60 sec
 ARIMA(1,0,1)(3,0,2)[7]             : OOB=1059411523.657, Time=1.65 sec
 ARIMA(1,0,1)(3,0,3)[7]             : OOB=962839692.586, Time=2.00 sec
 ARIMA(1,0,1)(4,0,0)[7]             : OOB=3563522122.458, Time=1.77 sec
 ARIMA(1,0,1)(4,0,1)[7]             : OOB=953388799.115, Time=2.70 sec
 ARIMA(1,0,1)(4,0,2)[7]             : OOB=928785114.567, Time=2.87 sec
 ARIMA(1,0,1)(5,0,0)[7]             : OOB=2403443268.018, Time=2.48 sec
 ARIMA(1,0,1)(5,0,1)[7]             : OOB=932157609.598, Time=4.95 sec
 ARIMA(1,0,2)(0,0,0)[7]             : OOB=6342022101.621, Time=54.77 sec
 ARIMA(1,0,2)(0,0,1)[7]             : OOB=2910307253.224, Time=0.34 sec
 ARIMA(1,0,2)(0,0,2)[7]             : OOB=1637796299.174, Time=0.75 sec
 ARIMA(1,0,2)(0,0,3)[7]             : OOB=2169016693.129, Time=1.63 sec
 ARIMA(1,0,2)(0,0,4)[7]             : OOB=3093100016.770, Time=2.75 sec
 ARIMA(1,0,2)(0,0,5)[7]             : OOB=5112027866.102, Time=4.72 sec
 ARIMA(1,0,2)(1,0,0)[7]             : OOB=9120519826.470, Time=0.50 sec
 ARIMA(1,0,2)(1,0,1)[7]             : OOB=2496891571.079, Time=0.73 sec
 ARIMA(1,0,2)(1,0,2)[7]             : OOB=977086347.407, Time=1.25 sec
 ARIMA(1,0,2)(1,0,3)[7]             : OOB=1425792069.242, Time=2.25 sec
 ARIMA(1,0,2)(1,0,4)[7]             : OOB=1330818514.644, Time=4.15 sec
 ARIMA(1,0,2)(2,0,0)[7]             : OOB=6794662444.588, Time=0.89 sec
 ARIMA(1,0,2)(2,0,1)[7]             : OOB=1416681853.191, Time=1.29 sec
 ARIMA(1,0,2)(2,0,2)[7]             : OOB=1099407772.213, Time=1.80 sec
 ARIMA(1,0,2)(2,0,3)[7]             : OOB=981585766.765, Time=3.12 sec
 ARIMA(1,0,2)(3,0,0)[7]             : OOB=5416086842.696, Time=1.47 sec
 ARIMA(1,0,2)(3,0,1)[7]             : OOB=1245256309.399, Time=2.63 sec
 ARIMA(1,0,2)(3,0,2)[7]             : OOB=1015868656.858, Time=3.78 sec
 ARIMA(1,0,2)(4,0,0)[7]             : OOB=3823099441.190, Time=2.53 sec
 ARIMA(1,0,2)(4,0,1)[7]             : OOB=910080591.741, Time=4.03 sec
 ARIMA(1,0,2)(5,0,0)[7]             : OOB=2564058649.852, Time=3.78 sec
 ARIMA(1,0,3)(0,0,0)[7]             : OOB=8997001101.819, Time=0.42 sec
 ARIMA(1,0,3)(0,0,1)[7]             : OOB=3920184537.202, Time=0.80 sec
 ARIMA(1,0,3)(0,0,2)[7]             : OOB=1977816513.064, Time=0.55 sec
 ARIMA(1,0,3)(0,0,3)[7]             : OOB=1407552621.608, Time=1.01 sec
 ARIMA(1,0,3)(0,0,4)[7]             : OOB=1779514020.289, Time=1.97 sec
 ARIMA(1,0,3)(1,0,0)[7]             : OOB=6798531319.209, Time=0.55 sec
 ARIMA(1,0,3)(1,0,1)[7]             : OOB=3014446667.832, Time=0.86 sec
 ARIMA(1,0,3)(1,0,2)[7]             : OOB=948149483.762, Time=1.66 sec
 ARIMA(1,0,3)(1,0,3)[7]             : OOB=1326347169.885, Time=3.00 sec
 ARIMA(1,0,3)(2,0,0)[7]             : OOB=7403811184.313, Time=0.85 sec
 ARIMA(1,0,3)(2,0,1)[7]             : OOB=1291866844.337, Time=1.42 sec
 ARIMA(1,0,3)(2,0,2)[7]             : OOB=995337464.566, Time=1.83 sec
 ARIMA(1,0,3)(3,0,0)[7]             : OOB=6161390503.788, Time=1.39 sec
 ARIMA(1,0,3)(3,0,1)[7]             : OOB=1120693649.488, Time=2.50 sec
 ARIMA(1,0,3)(4,0,0)[7]             : OOB=4482677219.493, Time=2.26 sec
 ARIMA(1,0,4)(0,0,0)[7]             : OOB=8277316945.642, Time=0.40 sec
 ARIMA(1,0,4)(0,0,1)[7]             : OOB=3981449269.363, Time=0.54 sec
 ARIMA(1,0,4)(0,0,2)[7]             : OOB=2374893698.265, Time=3.05 sec
 ARIMA(1,0,4)(0,0,3)[7]             : OOB=1408406952.833, Time=1.73 sec
 ARIMA(1,0,4)(1,0,0)[7]             : OOB=6765944848.567, Time=0.46 sec
 ARIMA(1,0,4)(1,0,1)[7]             : OOB=2855723802.504, Time=1.01 sec
 ARIMA(1,0,4)(1,0,2)[7]             : OOB=911181081.236, Time=2.27 sec
 ARIMA(1,0,4)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(1,0,4)(2,0,1)[7]             : OOB=1215763020.828, Time=1.58 sec
 ARIMA(1,0,4)(3,0,0)[7]             : OOB=6161434721.939, Time=1.59 sec
 ARIMA(1,0,5)(0,0,0)[7]             : OOB=1954199060.246, Time=0.46 sec
 ARIMA(1,0,5)(0,0,1)[7]             : OOB=1897780827.163, Time=0.73 sec
 ARIMA(1,0,5)(0,0,2)[7]             : OOB=1826766029.429, Time=1.46 sec
 ARIMA(1,0,5)(1,0,0)[7]             : OOB=6649753421.062, Time=0.51 sec
 ARIMA(1,0,5)(1,0,1)[7]             : OOB=2843797342.891, Time=1.35 sec
 ARIMA(1,0,5)(2,0,0)[7]             : OOB=7466689973.733, Time=1.11 sec
 ARIMA(2,0,0)(0,0,0)[7]             : OOB=7050535117.618, Time=0.09 sec
 ARIMA(2,0,0)(0,0,1)[7]             : OOB=7338135974.816, Time=0.22 sec
 ARIMA(2,0,0)(0,0,2)[7]             : OOB=9280760940.043, Time=0.50 sec
 ARIMA(2,0,0)(0,0,3)[7]             : OOB=11137496060.108, Time=1.12 sec
 ARIMA(2,0,0)(0,0,4)[7]             : OOB=12810292577.367, Time=2.18 sec
 ARIMA(2,0,0)(0,0,5)[7]             : OOB=13037003044.553, Time=3.93 sec
 ARIMA(2,0,0)(1,0,0)[7]             : OOB=8743769000.852, Time=0.34 sec
 ARIMA(2,0,0)(1,0,1)[7]             : OOB=inf, Time=nan sec
 ARIMA(2,0,0)(1,0,2)[7]             : OOB=2273147846.240, Time=1.13 sec
 ARIMA(2,0,0)(1,0,3)[7]             : OOB=3103828807.304, Time=1.88 sec
 ARIMA(2,0,0)(1,0,4)[7]             : OOB=2993414773.240, Time=3.83 sec
 ARIMA(2,0,0)(1,0,5)[7]             : OOB=2742198922.226, Time=6.62 sec
 ARIMA(2,0,0)(2,0,0)[7]             : OOB=4246525921.947, Time=1.13 sec
 ARIMA(2,0,0)(2,0,1)[7]             : OOB=3072099379.722, Time=0.96 sec
 ARIMA(2,0,0)(2,0,2)[7]             : OOB=2739197998.386, Time=1.29 sec
 ARIMA(2,0,0)(2,0,3)[7]             : OOB=2394995694.062, Time=1.67 sec
 ARIMA(2,0,0)(2,0,4)[7]             : OOB=2231332764.651, Time=3.88 sec
 ARIMA(2,0,0)(3,0,0)[7]             : OOB=2418226096.444, Time=1.98 sec
 ARIMA(2,0,0)(3,0,1)[7]             : OOB=2877448890.931, Time=1.98 sec
 ARIMA(2,0,0)(3,0,2)[7]             : OOB=2510906131.163, Time=2.10 sec
 ARIMA(2,0,0)(3,0,3)[7]             : OOB=inf, Time=3.69 sec
 ARIMA(2,0,0)(4,0,0)[7]             : OOB=inf, Time=1.92 sec
 ARIMA(2,0,0)(4,0,1)[7]             : OOB=2303809485.912, Time=3.12 sec
 ARIMA(2,0,0)(4,0,2)[7]             : OOB=2251182789.669, Time=4.23 sec
 ARIMA(2,0,0)(5,0,0)[7]             : OOB=inf, Time=3.33 sec
 ARIMA(2,0,0)(5,0,1)[7]             : OOB=2268362610.969, Time=7.83 sec
 ARIMA(2,0,1)(0,0,0)[7]             : OOB=6247450595.262, Time=0.41 sec
 ARIMA(2,0,1)(0,0,1)[7]             : OOB=8142430135.702, Time=0.64 sec
 ARIMA(2,0,1)(0,0,2)[7]             : OOB=12069899805.968, Time=0.80 sec
 ARIMA(2,0,1)(0,0,3)[7]             : OOB=14373947384.853, Time=1.52 sec
 ARIMA(2,0,1)(0,0,4)[7]             : OOB=16054481087.576, Time=2.37 sec
 ARIMA(2,0,1)(0,0,5)[7]             : OOB=16450603731.122, Time=5.32 sec
 ARIMA(2,0,1)(1,0,0)[7]             : OOB=8320464149.425, Time=1.02 sec
 ARIMA(2,0,1)(1,0,1)[7]             : OOB=2595491488.416, Time=0.51 sec
 ARIMA(2,0,1)(1,0,2)[7]             : OOB=965648850.081, Time=1.49 sec
 ARIMA(2,0,1)(1,0,3)[7]             : OOB=1393873614.029, Time=2.45 sec
 ARIMA(2,0,1)(1,0,4)[7]             : OOB=1289618269.827, Time=3.58 sec
 ARIMA(2,0,1)(2,0,0)[7]             : OOB=7011011526.822, Time=1.70 sec
 ARIMA(2,0,1)(2,0,1)[7]             : OOB=1380203344.759, Time=1.65 sec
 ARIMA(2,0,1)(2,0,2)[7]             : OOB=1056634202.212, Time=1.40 sec
 ARIMA(2,0,1)(2,0,3)[7]             : OOB=932695180.764, Time=2.03 sec
 ARIMA(2,0,1)(3,0,0)[7]             : OOB=5598255254.315, Time=1.38 sec
 ARIMA(2,0,1)(3,0,1)[7]             : OOB=1202020213.697, Time=2.31 sec
 ARIMA(2,0,1)(3,0,2)[7]             : OOB=975340335.139, Time=2.27 sec
 ARIMA(2,0,1)(4,0,0)[7]             : OOB=3937290104.746, Time=2.37 sec
 ARIMA(2,0,1)(4,0,1)[7]             : OOB=880323917.165, Time=3.52 sec
 ARIMA(2,0,1)(5,0,0)[7]             : OOB=2638523173.987, Time=5.45 sec
 ARIMA(2,0,2)(0,0,0)[7]             : OOB=7726502824.072, Time=0.36 sec
 ARIMA(2,0,2)(0,0,1)[7]             : OOB=8535962150.350, Time=1.01 sec
 ARIMA(2,0,2)(0,0,2)[7]             : OOB=8928823238.633, Time=0.99 sec
 ARIMA(2,0,2)(0,0,3)[7]             : OOB=9184627621.294, Time=1.90 sec
 ARIMA(2,0,2)(0,0,4)[7]             : OOB=9403425590.902, Time=3.83 sec
 ARIMA(2,0,2)(1,0,0)[7]             : OOB=9181426863.823, Time=0.69 sec
 ARIMA(2,0,2)(1,0,1)[7]             : OOB=2333926223.832, Time=0.66 sec
 ARIMA(2,0,2)(1,0,2)[7]             : OOB=1336899434.218, Time=1.43 sec
 ARIMA(2,0,2)(1,0,3)[7]             : OOB=1767952208.647, Time=7.81 sec
 ARIMA(2,0,2)(2,0,0)[7]             : OOB=6162668474.877, Time=1.63 sec
 ARIMA(2,0,2)(2,0,1)[7]             : OOB=1763505747.315, Time=1.23 sec
 ARIMA(2,0,2)(2,0,2)[7]             : OOB=1792774112.252, Time=1.68 sec
 ARIMA(2,0,2)(3,0,0)[7]             : OOB=4461525470.034, Time=2.85 sec
 ARIMA(2,0,2)(3,0,1)[7]             : OOB=1784654753.630, Time=2.50 sec
 ARIMA(2,0,2)(4,0,0)[7]             : OOB=3231441141.397, Time=4.76 sec
 ARIMA(2,0,3)(0,0,0)[7]             : OOB=5142409873.176, Time=0.66 sec
 ARIMA(2,0,3)(0,0,1)[7]             : OOB=4115544902.569, Time=2.37 sec
 ARIMA(2,0,3)(0,0,2)[7]             : OOB=2350937100.433, Time=1.31 sec
 ARIMA(2,0,3)(0,0,3)[7]             : OOB=1414216377.241, Time=1.59 sec
 ARIMA(2,0,3)(1,0,0)[7]             : OOB=6921558812.201, Time=0.67 sec
 ARIMA(2,0,3)(1,0,1)[7]             : OOB=2858113256.414, Time=1.10 sec
 ARIMA(2,0,3)(1,0,2)[7]             : OOB=1494765790.969, Time=2.35 sec
 ARIMA(2,0,3)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(2,0,3)(2,0,1)[7]             : OOB=1021149104.433, Time=1.94 sec
 ARIMA(2,0,3)(3,0,0)[7]             : OOB=4789909987.367, Time=2.40 sec
 ARIMA(2,0,4)(0,0,0)[7]             : OOB=1970324943.558, Time=1.49 sec
 ARIMA(2,0,4)(0,0,1)[7]             : OOB=4584316007.375, Time=1.18 sec
 ARIMA(2,0,4)(0,0,2)[7]             : OOB=2254276049.629, Time=1.85 sec
 ARIMA(2,0,4)(1,0,0)[7]             : OOB=6799093990.307, Time=1.74 sec
 ARIMA(2,0,4)(1,0,1)[7]             : OOB=1204932158.184, Time=2.12 sec
 ARIMA(2,0,4)(2,0,0)[7]             : OOB=7563067122.782, Time=1.88 sec
 ARIMA(2,0,5)(0,0,0)[7]             : OOB=1890169986.633, Time=1.45 sec
 ARIMA(2,0,5)(0,0,1)[7]             : OOB=1770576714.777, Time=2.18 sec
 ARIMA(2,0,5)(1,0,0)[7]             : OOB=6843277574.520, Time=0.56 sec
 ARIMA(3,0,0)(0,0,0)[7]             : OOB=inf, Time=0.57 sec
 ARIMA(3,0,0)(0,0,1)[7]             : OOB=6398905019.144, Time=2.37 sec
 ARIMA(3,0,0)(0,0,2)[7]             : OOB=7847017392.862, Time=0.90 sec
 ARIMA(3,0,0)(0,0,3)[7]             : OOB=10309490789.716, Time=1.19 sec
 ARIMA(3,0,0)(0,0,4)[7]             : OOB=11289642308.883, Time=2.49 sec
 ARIMA(3,0,0)(0,0,5)[7]             : OOB=11419481845.403, Time=5.87 sec
 ARIMA(3,0,0)(1,0,0)[7]             : OOB=8650598608.620, Time=0.38 sec
 ARIMA(3,0,0)(1,0,1)[7]             : OOB=1232391590.953, Time=0.63 sec
 ARIMA(3,0,0)(1,0,2)[7]             : OOB=1660630976.004, Time=1.12 sec
 ARIMA(3,0,0)(1,0,3)[7]             : OOB=2457658865.003, Time=2.20 sec
 ARIMA(3,0,0)(1,0,4)[7]             : OOB=2410306487.191, Time=3.57 sec
 ARIMA(3,0,0)(2,0,0)[7]             : OOB=4610499852.644, Time=0.74 sec
 ARIMA(3,0,0)(2,0,1)[7]             : OOB=2473593337.618, Time=1.24 sec
 ARIMA(3,0,0)(2,0,2)[7]             : OOB=2103040832.244, Time=1.35 sec
 ARIMA(3,0,0)(2,0,3)[7]             : OOB=1851754995.117, Time=2.93 sec
 ARIMA(3,0,0)(3,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(3,0,0)(3,0,1)[7]             : OOB=2276407452.333, Time=2.36 sec
 ARIMA(3,0,0)(3,0,2)[7]             : OOB=1926357791.534, Time=2.69 sec
 ARIMA(3,0,0)(4,0,0)[7]             : OOB=inf, Time=2.74 sec
 ARIMA(3,0,0)(4,0,1)[7]             : OOB=1785495642.261, Time=4.79 sec
 ARIMA(3,0,0)(5,0,0)[7]             : OOB=inf, Time=4.37 sec
 ARIMA(3,0,1)(0,0,0)[7]             : OOB=8454711605.311, Time=0.36 sec
 ARIMA(3,0,1)(0,0,1)[7]             : OOB=8635346411.669, Time=0.85 sec
 ARIMA(3,0,1)(0,0,2)[7]             : OOB=2165886636.520, Time=1.09 sec
 ARIMA(3,0,1)(0,0,3)[7]             : OOB=1420075256.176, Time=2.31 sec
 ARIMA(3,0,1)(0,0,4)[7]             : OOB=2082815261.928, Time=3.38 sec
 ARIMA(3,0,1)(1,0,0)[7]             : OOB=6032747659.461, Time=1.72 sec
 ARIMA(3,0,1)(1,0,1)[7]             : OOB=2757297274.960, Time=1.00 sec
 ARIMA(3,0,1)(1,0,2)[7]             : OOB=972754549.821, Time=1.16 sec
 ARIMA(3,0,1)(1,0,3)[7]             : OOB=1403509740.948, Time=2.87 sec
 ARIMA(3,0,1)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(3,0,1)(2,0,1)[7]             : OOB=1380727058.840, Time=2.23 sec
 ARIMA(3,0,1)(2,0,2)[7]             : OOB=1070035953.731, Time=1.88 sec
 ARIMA(3,0,1)(3,0,0)[7]             : OOB=5759626729.042, Time=2.00 sec
 ARIMA(3,0,1)(3,0,1)[7]             : OOB=1203040676.829, Time=2.46 sec
 ARIMA(3,0,1)(4,0,0)[7]             : OOB=4123921796.743, Time=3.01 sec
 ARIMA(3,0,2)(0,0,0)[7]             : OOB=5322037596.962, Time=0.43 sec
 ARIMA(3,0,2)(0,0,1)[7]             : OOB=8513131397.386, Time=1.76 sec
 ARIMA(3,0,2)(0,0,2)[7]             : OOB=2355895822.201, Time=2.44 sec
 ARIMA(3,0,2)(0,0,3)[7]             : OOB=1359344581.760, Time=2.89 sec
 ARIMA(3,0,2)(1,0,0)[7]             : OOB=7143556706.350, Time=1.33 sec
 ARIMA(3,0,2)(1,0,1)[7]             : OOB=3467039916.540, Time=1.22 sec
 ARIMA(3,0,2)(1,0,2)[7]             : OOB=1088586213.450, Time=1.81 sec
 ARIMA(3,0,2)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(3,0,2)(2,0,1)[7]             : OOB=1345781239.036, Time=1.76 sec
 ARIMA(3,0,2)(3,0,0)[7]             : OOB=6203421906.617, Time=3.03 sec
 ARIMA(3,0,3)(0,0,0)[7]             : OOB=7295476172.339, Time=1.21 sec
 ARIMA(3,0,3)(0,0,1)[7]             : OOB=3069315265.655, Time=3.82 sec
 ARIMA(3,0,3)(0,0,2)[7]             : OOB=1924429111.053, Time=6.90 sec
 ARIMA(3,0,3)(1,0,0)[7]             : OOB=1212068987.032, Time=1.07 sec
 ARIMA(3,0,3)(1,0,1)[7]             : OOB=1099978762.503, Time=4.66 sec
 ARIMA(3,0,3)(2,0,0)[7]             : OOB=5799440549.037, Time=2.03 sec
 ARIMA(3,0,4)(0,0,0)[7]             : OOB=7145720080.381, Time=1.99 sec
 ARIMA(3,0,4)(0,0,1)[7]             : OOB=2684390826.299, Time=5.71 sec
 ARIMA(3,0,4)(1,0,0)[7]             : OOB=1215070245.231, Time=1.36 sec
 ARIMA(3,0,5)(0,0,0)[7]             : OOB=5016509859.042, Time=1.93 sec
 ARIMA(4,0,0)(0,0,0)[7]             : OOB=inf, Time=0.95 sec
 ARIMA(4,0,0)(0,0,1)[7]             : OOB=2405436596.582, Time=0.87 sec
 ARIMA(4,0,0)(0,0,2)[7]             : OOB=3557379028.454, Time=0.74 sec
 ARIMA(4,0,0)(0,0,3)[7]             : OOB=5986827722.661, Time=2.09 sec
 ARIMA(4,0,0)(0,0,4)[7]             : OOB=7296197047.541, Time=3.48 sec
 ARIMA(4,0,0)(1,0,0)[7]             : OOB=9908949392.116, Time=3.92 sec
 ARIMA(4,0,0)(1,0,1)[7]             : OOB=2234482293.384, Time=2.12 sec
 ARIMA(4,0,0)(1,0,2)[7]             : OOB=1072831878.104, Time=1.37 sec
 ARIMA(4,0,0)(1,0,3)[7]             : OOB=1536842720.777, Time=3.69 sec
 ARIMA(4,0,0)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(4,0,0)(2,0,1)[7]             : OOB=1528150033.746, Time=1.68 sec
 ARIMA(4,0,0)(2,0,2)[7]             : OOB=1347034113.308, Time=1.95 sec
 ARIMA(4,0,0)(3,0,0)[7]             : OOB=4939067646.895, Time=1.97 sec
 ARIMA(4,0,0)(3,0,1)[7]             : OOB=1425659835.871, Time=3.27 sec
 ARIMA(4,0,0)(4,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(4,0,1)(0,0,0)[7]             : OOB=7507158854.598, Time=0.95 sec
 ARIMA(4,0,1)(0,0,1)[7]             : OOB=3090074323.239, Time=1.13 sec
 ARIMA(4,0,1)(0,0,2)[7]             : OOB=1826382804.722, Time=1.43 sec
 ARIMA(4,0,1)(0,0,3)[7]             : OOB=1426856534.321, Time=1.98 sec
 ARIMA(4,0,1)(1,0,0)[7]             : OOB=7248264714.926, Time=2.11 sec
 ARIMA(4,0,1)(1,0,1)[7]             : OOB=3097639050.831, Time=0.98 sec
 ARIMA(4,0,1)(1,0,2)[7]             : OOB=957722078.983, Time=1.43 sec
 ARIMA(4,0,1)(2,0,0)[7]             : OOB=7622244538.962, Time=1.50 sec
 ARIMA(4,0,1)(2,0,1)[7]             : OOB=1282212419.988, Time=3.03 sec
 ARIMA(4,0,1)(3,0,0)[7]             : OOB=6361175705.857, Time=2.52 sec
 ARIMA(4,0,2)(0,0,0)[7]             : OOB=4341010581.760, Time=0.62 sec
 ARIMA(4,0,2)(0,0,1)[7]             : OOB=3383878542.895, Time=1.11 sec
 ARIMA(4,0,2)(0,0,2)[7]             : OOB=9194845092.347, Time=2.43 sec
 ARIMA(4,0,2)(1,0,0)[7]             : OOB=7358604894.647, Time=0.98 sec
 ARIMA(4,0,2)(1,0,1)[7]             : OOB=3396139411.408, Time=1.45 sec
 ARIMA(4,0,2)(2,0,0)[7]             : OOB=inf, Time=2.96 sec
 ARIMA(4,0,3)(0,0,0)[7]             : OOB=5061837080.405, Time=0.73 sec
 ARIMA(4,0,3)(0,0,1)[7]             : OOB=inf, Time=5.56 sec
 ARIMA(4,0,3)(1,0,0)[7]             : OOB=inf, Time=2.43 sec
 ARIMA(4,0,4)(0,0,0)[7]             : OOB=7454171185.866, Time=1.20 sec
 ARIMA(5,0,0)(0,0,0)[7]             : OOB=inf, Time=0.20 sec
 ARIMA(5,0,0)(0,0,1)[7]             : OOB=2532224923.201, Time=0.52 sec
 ARIMA(5,0,0)(0,0,2)[7]             : OOB=1644690467.344, Time=0.49 sec
 ARIMA(5,0,0)(0,0,3)[7]             : OOB=1507733643.181, Time=1.39 sec
 ARIMA(5,0,0)(1,0,0)[7]             : OOB=9821601127.039, Time=0.83 sec
 ARIMA(5,0,0)(1,0,1)[7]             : OOB=2330169110.132, Time=1.12 sec
 ARIMA(5,0,0)(1,0,2)[7]             : OOB=1123844588.288, Time=2.73 sec
 ARIMA(5,0,0)(2,0,0)[7]             : OOB=6693558965.863, Time=2.81 sec
 ARIMA(5,0,0)(2,0,1)[7]             : OOB=1576170842.432, Time=3.35 sec
 ARIMA(5,0,0)(3,0,0)[7]             : OOB=5030867739.901, Time=4.69 sec
 ARIMA(5,0,1)(0,0,0)[7]             : OOB=7548287927.500, Time=0.46 sec
 ARIMA(5,0,1)(0,0,1)[7]             : OOB=3166167063.870, Time=0.97 sec
 ARIMA(5,0,1)(0,0,2)[7]             : OOB=1849007593.256, Time=1.00 sec
 ARIMA(5,0,1)(1,0,0)[7]             : OOB=5814952065.506, Time=1.13 sec
 ARIMA(5,0,1)(1,0,1)[7]             : OOB=3018659415.993, Time=0.95 sec
 ARIMA(5,0,1)(2,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(5,0,2)(0,0,0)[7]             : OOB=7220360647.869, Time=1.52 sec
 ARIMA(5,0,2)(0,0,1)[7]             : OOB=3121862613.198, Time=1.26 sec
 ARIMA(5,0,2)(1,0,0)[7]             : OOB=6279899933.483, Time=0.93 sec
 ARIMA(5,0,3)(0,0,0)[7]             : OOB=inf, Time=2.23 sec

Best model:  ARIMA(0,0,0)(5,0,1)[7]          
Total fit time: 1089.639 seconds

ARIMA(maxiter=200, order=(0, 0, 0), out_of_sample_size=30, scoring_args={},
      seasonal_order=(5, 0, 1, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((0,0,0),(5,0,1,7)), aic_bic=False).rmse()

25066.46099110815


#Might take some time to finish

auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=0, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False, stepwise=False, max_order=8)

 ARIMA(0,1,0)(0,0,0)[7]             : OOB=2907757359.033, Time=0.34 sec
 ARIMA(0,1,0)(0,0,1)[7]             : OOB=2366460868.806, Time=0.23 sec
 ARIMA(0,1,0)(0,0,2)[7]             : OOB=2522191770.201, Time=0.37 sec
 ARIMA(0,1,0)(0,0,3)[7]             : OOB=3323220171.972, Time=0.57 sec
 ARIMA(0,1,0)(0,0,4)[7]             : OOB=4624492301.100, Time=1.01 sec
 ARIMA(0,1,0)(0,0,5)[7]             : OOB=4792763473.571, Time=1.56 sec
 ARIMA(0,1,0)(1,0,0)[7]             : OOB=5723066035.047, Time=0.10 sec
 ARIMA(0,1,0)(1,0,1)[7]             : OOB=5048755556.445, Time=0.21 sec
 ARIMA(0,1,0)(1,0,2)[7]             : OOB=1388622534.486, Time=0.53 sec
 ARIMA(0,1,0)(1,0,3)[7]             : OOB=1262062747.826, Time=0.90 sec
 ARIMA(0,1,0)(1,0,4)[7]             : OOB=1122647948.621, Time=1.44 sec
 ARIMA(0,1,0)(1,0,5)[7]             : OOB=1063867969.612, Time=3.00 sec
 ARIMA(0,1,0)(2,0,0)[7]             : OOB=8687011702.573, Time=0.25 sec
 ARIMA(0,1,0)(2,0,1)[7]             : OOB=1268776359.087, Time=0.40 sec
 ARIMA(0,1,0)(2,0,2)[7]             : OOB=1131044120.197, Time=0.72 sec
 ARIMA(0,1,0)(2,0,3)[7]             : OOB=1120801118.208, Time=1.69 sec
 ARIMA(0,1,0)(2,0,4)[7]             : OOB=1200160667.520, Time=2.06 sec
 ARIMA(0,1,0)(2,0,5)[7]             : OOB=1169810123.516, Time=3.78 sec
 ARIMA(0,1,0)(3,0,0)[7]             : OOB=10095174215.807, Time=81.97 sec
 ARIMA(0,1,0)(3,0,1)[7]             : OOB=1148778749.150, Time=0.94 sec
 ARIMA(0,1,0)(3,0,2)[7]             : OOB=1272533089.151, Time=0.92 sec
 ARIMA(0,1,0)(3,0,3)[7]             : OOB=1112804709.462, Time=2.12 sec
 ARIMA(0,1,0)(3,0,4)[7]             : OOB=1216095694.917, Time=2.27 sec
 ARIMA(0,1,0)(3,0,5)[7]             : OOB=1186381481.234, Time=4.14 sec
 ARIMA(0,1,0)(4,0,0)[7]             : OOB=9859958883.437, Time=0.91 sec
 ARIMA(0,1,0)(4,0,1)[7]             : OOB=1124561219.249, Time=1.67 sec
 ARIMA(0,1,0)(4,0,2)[7]             : OOB=1093749243.679, Time=2.24 sec
 ARIMA(0,1,0)(4,0,3)[7]             : OOB=inf, Time=5.33 sec
 ARIMA(0,1,0)(4,0,4)[7]             : OOB=inf, Time=5.27 sec
 ARIMA(0,1,0)(5,0,0)[7]             : OOB=inf, Time=1.71 sec
 ARIMA(0,1,0)(5,0,1)[7]             : OOB=1118962391.040, Time=3.55 sec
 ARIMA(0,1,0)(5,0,2)[7]             : OOB=1125543029.020, Time=4.59 sec
 ARIMA(0,1,0)(5,0,3)[7]             : OOB=inf, Time=10.46 sec
 ARIMA(0,1,1)(0,0,0)[7]             : OOB=2900216712.111, Time=0.07 sec
 ARIMA(0,1,1)(0,0,1)[7]             : OOB=2350368698.239, Time=0.16 sec
 ARIMA(0,1,1)(0,0,2)[7]             : OOB=2185451296.647, Time=0.34 sec
 ARIMA(0,1,1)(0,0,3)[7]             : OOB=2652430604.514, Time=0.58 sec
 ARIMA(0,1,1)(0,0,4)[7]             : OOB=3332884715.906, Time=1.04 sec
 ARIMA(0,1,1)(0,0,5)[7]             : OOB=4051872458.251, Time=1.88 sec
 ARIMA(0,1,1)(1,0,0)[7]             : OOB=8089722602.391, Time=0.18 sec
 ARIMA(0,1,1)(1,0,1)[7]             : OOB=9397316159.953, Time=0.32 sec
 ARIMA(0,1,1)(1,0,2)[7]             : OOB=3369948122.371, Time=0.52 sec
 ARIMA(0,1,1)(1,0,3)[7]             : OOB=1953348658.492, Time=0.87 sec
 ARIMA(0,1,1)(1,0,4)[7]             : OOB=1655834724.898, Time=1.65 sec
 ARIMA(0,1,1)(1,0,5)[7]             : OOB=1730883671.167, Time=2.97 sec
 ARIMA(0,1,1)(2,0,0)[7]             : OOB=10398117270.713, Time=0.29 sec
 ARIMA(0,1,1)(2,0,1)[7]             : OOB=1661380113.665, Time=0.52 sec
 ARIMA(0,1,1)(2,0,2)[7]             : OOB=1710915753.997, Time=0.82 sec
 ARIMA(0,1,1)(2,0,3)[7]             : OOB=2067141706.594, Time=1.71 sec
 ARIMA(0,1,1)(2,0,4)[7]             : OOB=1690403002.766, Time=1.49 sec
 ARIMA(0,1,1)(2,0,5)[7]             : OOB=1667092507.939, Time=4.22 sec
 ARIMA(0,1,1)(3,0,0)[7]             : OOB=12668498805.449, Time=0.62 sec
 ARIMA(0,1,1)(3,0,1)[7]             : OOB=1621093623.919, Time=1.11 sec
 ARIMA(0,1,1)(3,0,2)[7]             : OOB=1728612607.538, Time=1.05 sec
 ARIMA(0,1,1)(3,0,3)[7]             : OOB=2036392684.651, Time=1.90 sec
 ARIMA(0,1,1)(3,0,4)[7]             : OOB=inf, Time=3.02 sec
 ARIMA(0,1,1)(4,0,0)[7]             : OOB=11722462556.394, Time=1.14 sec
 ARIMA(0,1,1)(4,0,1)[7]             : OOB=2091770248.399, Time=2.05 sec
 ARIMA(0,1,1)(4,0,2)[7]             : OOB=1767471091.181, Time=2.72 sec
 ARIMA(0,1,1)(4,0,3)[7]             : OOB=1627949772.864, Time=2.77 sec
 ARIMA(0,1,1)(5,0,0)[7]             : OOB=10059888828.477, Time=2.07 sec
 ARIMA(0,1,1)(5,0,1)[7]             : OOB=2246723839.760, Time=4.43 sec
 ARIMA(0,1,1)(5,0,2)[7]             : OOB=1974431064.563, Time=4.74 sec
 ARIMA(0,1,2)(0,0,0)[7]             : OOB=8163948787.149, Time=0.13 sec
 ARIMA(0,1,2)(0,0,1)[7]             : OOB=3665144518.596, Time=0.69 sec
 ARIMA(0,1,2)(0,0,2)[7]             : OOB=2183294342.824, Time=0.42 sec
 ARIMA(0,1,2)(0,0,3)[7]             : OOB=1352843636.830, Time=0.86 sec
 ARIMA(0,1,2)(0,0,4)[7]             : OOB=1663294674.031, Time=1.73 sec
 ARIMA(0,1,2)(0,0,5)[7]             : OOB=2751638511.539, Time=3.51 sec
 ARIMA(0,1,2)(1,0,0)[7]             : OOB=7253724931.199, Time=0.20 sec
 ARIMA(0,1,2)(1,0,1)[7]             : OOB=9299903114.300, Time=1.27 sec
 ARIMA(0,1,2)(1,0,2)[7]             : OOB=3233796554.072, Time=0.65 sec
 ARIMA(0,1,2)(1,0,3)[7]             : OOB=1806557842.983, Time=1.44 sec
 ARIMA(0,1,2)(1,0,4)[7]             : OOB=1532709879.825, Time=2.43 sec
 ARIMA(0,1,2)(1,0,5)[7]             : OOB=1609198713.092, Time=4.96 sec
 ARIMA(0,1,2)(2,0,0)[7]             : OOB=9969664921.845, Time=0.37 sec
 ARIMA(0,1,2)(2,0,1)[7]             : OOB=1555279001.027, Time=0.68 sec
 ARIMA(0,1,2)(2,0,2)[7]             : OOB=1628123811.506, Time=1.03 sec
 ARIMA(0,1,2)(2,0,3)[7]             : OOB=1932881614.862, Time=2.21 sec
 ARIMA(0,1,2)(2,0,4)[7]             : OOB=1572040674.217, Time=2.52 sec
 ARIMA(0,1,2)(3,0,0)[7]             : OOB=12572090461.394, Time=0.74 sec
 ARIMA(0,1,2)(3,0,1)[7]             : OOB=1516356507.218, Time=1.31 sec
 ARIMA(0,1,2)(3,0,2)[7]             : OOB=1609022885.865, Time=1.16 sec
 ARIMA(0,1,2)(3,0,3)[7]             : OOB=1930695391.088, Time=3.17 sec
 ARIMA(0,1,2)(4,0,0)[7]             : OOB=11758194894.896, Time=1.29 sec
 ARIMA(0,1,2)(4,0,1)[7]             : OOB=1978896019.703, Time=2.21 sec
 ARIMA(0,1,2)(4,0,2)[7]             : OOB=1661597110.964, Time=2.96 sec
 ARIMA(0,1,2)(5,0,0)[7]             : OOB=10157195368.440, Time=2.27 sec
 ARIMA(0,1,2)(5,0,1)[7]             : OOB=2100629708.878, Time=5.14 sec
 ARIMA(0,1,3)(0,0,0)[7]             : OOB=11549456292.573, Time=0.19 sec
 ARIMA(0,1,3)(0,0,1)[7]             : OOB=5727053143.425, Time=0.31 sec
 ARIMA(0,1,3)(0,0,2)[7]             : OOB=3312852835.611, Time=0.72 sec
 ARIMA(0,1,3)(0,0,3)[7]             : OOB=1489695440.909, Time=0.98 sec
 ARIMA(0,1,3)(0,0,4)[7]             : OOB=1300026945.265, Time=2.04 sec
 ARIMA(0,1,3)(0,0,5)[7]             : OOB=2167577264.144, Time=2.56 sec
 ARIMA(0,1,3)(1,0,0)[7]             : OOB=4626794121.415, Time=0.26 sec
 ARIMA(0,1,3)(1,0,1)[7]             : OOB=9018770858.697, Time=1.04 sec
 ARIMA(0,1,3)(1,0,2)[7]             : OOB=2937467101.953, Time=0.90 sec
 ARIMA(0,1,3)(1,0,3)[7]             : OOB=1573633365.712, Time=1.48 sec
 ARIMA(0,1,3)(1,0,4)[7]             : OOB=1278501146.951, Time=3.39 sec
 ARIMA(0,1,3)(2,0,0)[7]             : OOB=8600912336.138, Time=0.44 sec
 ARIMA(0,1,3)(2,0,1)[7]             : OOB=1297797117.472, Time=0.74 sec
 ARIMA(0,1,3)(2,0,2)[7]             : OOB=1391591574.468, Time=1.27 sec
 ARIMA(0,1,3)(2,0,3)[7]             : OOB=1672768431.463, Time=2.63 sec
 ARIMA(0,1,3)(3,0,0)[7]             : OOB=11366262467.288, Time=0.84 sec
 ARIMA(0,1,3)(3,0,1)[7]             : OOB=1297860457.083, Time=1.53 sec
 ARIMA(0,1,3)(3,0,2)[7]             : OOB=1340046314.066, Time=1.35 sec
 ARIMA(0,1,3)(4,0,0)[7]             : OOB=10934642489.336, Time=1.61 sec
 ARIMA(0,1,3)(4,0,1)[7]             : OOB=1746222900.266, Time=2.72 sec
 ARIMA(0,1,3)(5,0,0)[7]             : OOB=9438120691.354, Time=2.61 sec
 ARIMA(0,1,4)(0,0,0)[7]             : OOB=10631976495.080, Time=0.25 sec
 ARIMA(0,1,4)(0,0,1)[7]             : OOB=5810251376.600, Time=0.54 sec
 ARIMA(0,1,4)(0,0,2)[7]             : OOB=3727677544.757, Time=0.68 sec
 ARIMA(0,1,4)(0,0,3)[7]             : OOB=1590611223.525, Time=1.52 sec
 ARIMA(0,1,4)(0,0,4)[7]             : OOB=1282265694.033, Time=2.43 sec
 ARIMA(0,1,4)(1,0,0)[7]             : OOB=4551423317.960, Time=0.30 sec
 ARIMA(0,1,4)(1,0,1)[7]             : OOB=9341587969.971, Time=0.55 sec
 ARIMA(0,1,4)(1,0,2)[7]             : OOB=2991350565.672, Time=1.05 sec
 ARIMA(0,1,4)(1,0,3)[7]             : OOB=1718883160.718, Time=1.88 sec
 ARIMA(0,1,4)(2,0,0)[7]             : OOB=8880526400.026, Time=0.51 sec
 ARIMA(0,1,4)(2,0,1)[7]             : OOB=1422644415.269, Time=0.92 sec
 ARIMA(0,1,4)(2,0,2)[7]             : OOB=1497790645.737, Time=1.55 sec
 ARIMA(0,1,4)(3,0,0)[7]             : OOB=12004917777.883, Time=1.05 sec
 ARIMA(0,1,4)(3,0,1)[7]             : OOB=1408737851.986, Time=1.76 sec
 ARIMA(0,1,4)(4,0,0)[7]             : OOB=11525061866.333, Time=1.73 sec
 ARIMA(0,1,5)(0,0,0)[7]             : OOB=2523922632.700, Time=0.64 sec
 ARIMA(0,1,5)(0,0,1)[7]             : OOB=2561875430.492, Time=0.58 sec
 ARIMA(0,1,5)(0,0,2)[7]             : OOB=2717995880.376, Time=1.05 sec
 ARIMA(0,1,5)(0,0,3)[7]             : OOB=1365150772.483, Time=1.71 sec
 ARIMA(0,1,5)(1,0,0)[7]             : OOB=4480408352.497, Time=0.41 sec
 ARIMA(0,1,5)(1,0,1)[7]             : OOB=9337043711.288, Time=0.65 sec
 ARIMA(0,1,5)(1,0,2)[7]             : OOB=2993438434.590, Time=1.39 sec
 ARIMA(0,1,5)(2,0,0)[7]             : OOB=8828048497.340, Time=0.79 sec
 ARIMA(0,1,5)(2,0,1)[7]             : OOB=1427161621.807, Time=0.97 sec
 ARIMA(0,1,5)(3,0,0)[7]             : OOB=11925211811.171, Time=1.17 sec
 ARIMA(1,1,0)(0,0,0)[7]             : OOB=2929927085.689, Time=0.06 sec
 ARIMA(1,1,0)(0,0,1)[7]             : OOB=2373118438.856, Time=0.14 sec
 ARIMA(1,1,0)(0,0,2)[7]             : OOB=2611551474.578, Time=0.24 sec
 ARIMA(1,1,0)(0,0,3)[7]             : OOB=3428048271.805, Time=0.63 sec
 ARIMA(1,1,0)(0,0,4)[7]             : OOB=4825024491.583, Time=1.06 sec
 ARIMA(1,1,0)(0,0,5)[7]             : OOB=5133091237.951, Time=2.22 sec
 ARIMA(1,1,0)(1,0,0)[7]             : OOB=8627755110.970, Time=0.29 sec
 ARIMA(1,1,0)(1,0,1)[7]             : OOB=8469386486.172, Time=0.34 sec
 ARIMA(1,1,0)(1,0,2)[7]             : OOB=2462155325.421, Time=0.56 sec
 ARIMA(1,1,0)(1,0,3)[7]             : OOB=1642027196.154, Time=2.21 sec
 ARIMA(1,1,0)(1,0,4)[7]             : OOB=1440254252.284, Time=1.63 sec
 ARIMA(1,1,0)(1,0,5)[7]             : OOB=1472614880.313, Time=3.25 sec
 ARIMA(1,1,0)(2,0,0)[7]             : OOB=10474647991.954, Time=0.34 sec
 ARIMA(1,1,0)(2,0,1)[7]             : OOB=1462498443.502, Time=0.59 sec
 ARIMA(1,1,0)(2,0,2)[7]             : OOB=1458937173.722, Time=0.98 sec
 ARIMA(1,1,0)(2,0,3)[7]             : OOB=1689631952.239, Time=1.99 sec
 ARIMA(1,1,0)(2,0,4)[7]             : OOB=1467618553.623, Time=2.64 sec
 ARIMA(1,1,0)(2,0,5)[7]             : OOB=1439760042.392, Time=3.82 sec
 ARIMA(1,1,0)(3,0,0)[7]             : OOB=11703641036.143, Time=0.85 sec
 ARIMA(1,1,0)(3,0,1)[7]             : OOB=1447040930.559, Time=1.39 sec
 ARIMA(1,1,0)(3,0,2)[7]             : OOB=1500104691.555, Time=1.47 sec
 ARIMA(1,1,0)(3,0,3)[7]             : OOB=1614837824.169, Time=2.06 sec
 ARIMA(1,1,0)(3,0,4)[7]             : OOB=1633342051.774, Time=4.12 sec
 ARIMA(1,1,0)(4,0,0)[7]             : OOB=inf, Time=1.30 sec
 ARIMA(1,1,0)(4,0,1)[7]             : OOB=1686049789.608, Time=2.00 sec
 ARIMA(1,1,0)(4,0,2)[7]             : OOB=1510181783.017, Time=2.85 sec
 ARIMA(1,1,0)(4,0,3)[7]             : OOB=1429138954.704, Time=2.83 sec
 ARIMA(1,1,0)(5,0,0)[7]             : OOB=inf, Time=3.01 sec
 ARIMA(1,1,0)(5,0,1)[7]             : OOB=1725948195.673, Time=4.74 sec
 ARIMA(1,1,0)(5,0,2)[7]             : OOB=1549079873.076, Time=5.59 sec
 ARIMA(1,1,1)(0,0,0)[7]             : OOB=10060497263.033, Time=0.33 sec
 ARIMA(1,1,1)(0,0,1)[7]             : OOB=5535040626.768, Time=0.56 sec
 ARIMA(1,1,1)(0,0,2)[7]             : OOB=3585417815.936, Time=0.83 sec
 ARIMA(1,1,1)(0,0,3)[7]             : OOB=1565834584.481, Time=0.99 sec
 ARIMA(1,1,1)(0,0,4)[7]             : OOB=1301517642.946, Time=1.67 sec
 ARIMA(1,1,1)(0,0,5)[7]             : OOB=2112130335.509, Time=3.74 sec
 ARIMA(1,1,1)(1,0,0)[7]             : OOB=5569156125.017, Time=0.46 sec
 ARIMA(1,1,1)(1,0,1)[7]             : OOB=9819649607.774, Time=1.48 sec
 ARIMA(1,1,1)(1,0,2)[7]             : OOB=3222881498.122, Time=2.15 sec
 ARIMA(1,1,1)(1,0,3)[7]             : OOB=1761678539.157, Time=1.28 sec
 ARIMA(1,1,1)(1,0,4)[7]             : OOB=1478496869.839, Time=4.33 sec
 ARIMA(1,1,1)(1,0,5)[7]             : OOB=1544032217.664, Time=5.18 sec
 ARIMA(1,1,1)(2,0,0)[7]             : OOB=9390050621.356, Time=1.05 sec
 ARIMA(1,1,1)(2,0,1)[7]             : OOB=1509922839.891, Time=1.19 sec
 ARIMA(1,1,1)(2,0,2)[7]             : OOB=1588811051.052, Time=3.98 sec
 ARIMA(1,1,1)(2,0,3)[7]             : OOB=1881472558.815, Time=3.25 sec
 ARIMA(1,1,1)(2,0,4)[7]             : OOB=1525919689.274, Time=3.45 sec
 ARIMA(1,1,1)(3,0,0)[7]             : OOB=12195028887.125, Time=3.33 sec
 ARIMA(1,1,1)(3,0,1)[7]             : OOB=1465005048.411, Time=2.41 sec
 ARIMA(1,1,1)(3,0,2)[7]             : OOB=1560155461.543, Time=3.06 sec
 ARIMA(1,1,1)(3,0,3)[7]             : OOB=1885082620.820, Time=3.22 sec
 ARIMA(1,1,1)(4,0,0)[7]             : OOB=11518774102.100, Time=9.35 sec
 ARIMA(1,1,1)(4,0,1)[7]             : OOB=1929462234.964, Time=3.63 sec
 ARIMA(1,1,1)(4,0,2)[7]             : OOB=1609488990.212, Time=4.07 sec
 ARIMA(1,1,1)(5,0,0)[7]             : OOB=9998179834.107, Time=7.05 sec
 ARIMA(1,1,1)(5,0,1)[7]             : OOB=2040334749.318, Time=5.31 sec
 ARIMA(1,1,2)(0,0,0)[7]             : OOB=11067246537.913, Time=0.96 sec
 ARIMA(1,1,2)(0,0,1)[7]             : OOB=5632139927.701, Time=3.18 sec
 ARIMA(1,1,2)(0,0,2)[7]             : OOB=3479772929.153, Time=1.89 sec
 ARIMA(1,1,2)(0,0,3)[7]             : OOB=1526918443.169, Time=2.67 sec
 ARIMA(1,1,2)(0,0,4)[7]             : OOB=1311774555.274, Time=5.92 sec
 ARIMA(1,1,2)(0,0,5)[7]             : OOB=2099429064.094, Time=7.69 sec
 ARIMA(1,1,2)(1,0,0)[7]             : OOB=812676025.821, Time=1.04 sec
 ARIMA(1,1,2)(1,0,1)[7]             : OOB=9513595839.785, Time=0.91 sec
 ARIMA(1,1,2)(1,0,2)[7]             : OOB=3912464066.274, Time=1.50 sec
 ARIMA(1,1,2)(1,0,3)[7]             : OOB=2175494758.332, Time=2.74 sec
 ARIMA(1,1,2)(1,0,4)[7]             : OOB=1568930609.216, Time=4.21 sec
 ARIMA(1,1,2)(2,0,0)[7]             : OOB=inf, Time=1.25 sec
 ARIMA(1,1,2)(2,0,1)[7]             : OOB=1645000009.041, Time=1.27 sec
 ARIMA(1,1,2)(2,0,2)[7]             : OOB=1497976457.401, Time=2.49 sec
 ARIMA(1,1,2)(2,0,3)[7]             : OOB=inf, Time=6.27 sec
 ARIMA(1,1,2)(3,0,0)[7]             : OOB=12740738351.559, Time=2.35 sec
 ARIMA(1,1,2)(3,0,1)[7]             : OOB=1524840172.047, Time=2.44 sec
 ARIMA(1,1,2)(3,0,2)[7]             : OOB=1591233461.014, Time=2.17 sec
 ARIMA(1,1,2)(4,0,0)[7]             : OOB=12064698034.342, Time=3.02 sec
 ARIMA(1,1,2)(4,0,1)[7]             : OOB=1913372300.552, Time=2.83 sec
 ARIMA(1,1,2)(5,0,0)[7]             : OOB=10464121077.279, Time=2.81 sec
 ARIMA(1,1,3)(0,0,0)[7]             : OOB=11108130819.360, Time=0.41 sec
 ARIMA(1,1,3)(0,0,1)[7]             : OOB=5771788735.276, Time=1.84 sec
 ARIMA(1,1,3)(0,0,2)[7]             : OOB=3544964224.113, Time=0.97 sec
 ARIMA(1,1,3)(0,0,3)[7]             : OOB=1546560300.385, Time=1.46 sec
 ARIMA(1,1,3)(0,0,4)[7]             : OOB=1290140619.684, Time=3.96 sec
 ARIMA(1,1,3)(1,0,0)[7]             : OOB=4693335462.154, Time=1.05 sec
 ARIMA(1,1,3)(1,0,1)[7]             : OOB=10139477418.260, Time=1.82 sec
 ARIMA(1,1,3)(1,0,2)[7]             : OOB=2507048186.212, Time=2.39 sec
 ARIMA(1,1,3)(1,0,3)[7]             : OOB=1696113611.155, Time=3.79 sec
 ARIMA(1,1,3)(2,0,0)[7]             : OOB=9419340086.984, Time=1.10 sec
 ARIMA(1,1,3)(2,0,1)[7]             : OOB=1317075437.351, Time=1.66 sec
 ARIMA(1,1,3)(2,0,2)[7]             : OOB=1309126674.616, Time=225.82 sec
 ARIMA(1,1,3)(3,0,0)[7]             : OOB=12426584723.575, Time=2.88 sec
 ARIMA(1,1,3)(3,0,1)[7]             : OOB=1311084636.614, Time=2.84 sec
 ARIMA(1,1,3)(4,0,0)[7]             : OOB=11842378923.991, Time=3.49 sec
 ARIMA(1,1,4)(0,0,0)[7]             : OOB=2170504980.656, Time=0.38 sec
 ARIMA(1,1,4)(0,0,1)[7]             : OOB=2298007816.156, Time=1.16 sec
 ARIMA(1,1,4)(0,0,2)[7]             : OOB=2672957014.400, Time=2.19 sec
 ARIMA(1,1,4)(0,0,3)[7]             : OOB=1380880021.789, Time=1.96 sec
 ARIMA(1,1,4)(1,0,0)[7]             : OOB=577469731.385, Time=0.90 sec
 ARIMA(1,1,4)(1,0,1)[7]             : OOB=9381468912.971, Time=0.94 sec
 ARIMA(1,1,4)(1,0,2)[7]             : OOB=3112287300.687, Time=1.78 sec
 ARIMA(1,1,4)(2,0,0)[7]             : OOB=9266499028.869, Time=0.66 sec
 ARIMA(1,1,4)(2,0,1)[7]             : OOB=1765801828.926, Time=1.32 sec
 ARIMA(1,1,4)(3,0,0)[7]             : OOB=12989743947.446, Time=1.79 sec
 ARIMA(1,1,5)(0,0,0)[7]             : OOB=2211174548.594, Time=1.91 sec
 ARIMA(1,1,5)(0,0,1)[7]             : OOB=1852521855.199, Time=2.12 sec
 ARIMA(1,1,5)(0,0,2)[7]             : OOB=1748091667.748, Time=1.27 sec
 ARIMA(1,1,5)(1,0,0)[7]             : OOB=557860696.978, Time=1.23 sec
 ARIMA(1,1,5)(1,0,1)[7]             : OOB=9238451411.679, Time=1.12 sec
 ARIMA(1,1,5)(2,0,0)[7]             : OOB=9254772365.118, Time=0.70 sec
 ARIMA(2,1,0)(0,0,0)[7]             : OOB=3142217254.986, Time=0.48 sec
 ARIMA(2,1,0)(0,0,1)[7]             : OOB=2273067578.419, Time=0.23 sec
 ARIMA(2,1,0)(0,0,2)[7]             : OOB=2322355656.242, Time=0.37 sec
 ARIMA(2,1,0)(0,0,3)[7]             : OOB=3280059375.764, Time=0.89 sec
 ARIMA(2,1,0)(0,0,4)[7]             : OOB=4409723021.174, Time=1.51 sec
 ARIMA(2,1,0)(0,0,5)[7]             : OOB=4661157424.949, Time=2.71 sec
 ARIMA(2,1,0)(1,0,0)[7]             : OOB=8901318044.166, Time=0.56 sec
 ARIMA(2,1,0)(1,0,1)[7]             : OOB=8698857748.843, Time=0.41 sec
 ARIMA(2,1,0)(1,0,2)[7]             : OOB=2916979843.110, Time=0.62 sec
 ARIMA(2,1,0)(1,0,3)[7]             : OOB=1817395488.761, Time=1.20 sec
 ARIMA(2,1,0)(1,0,4)[7]             : OOB=1593988439.892, Time=1.97 sec
 ARIMA(2,1,0)(1,0,5)[7]             : OOB=1647440837.294, Time=3.68 sec
 ARIMA(2,1,0)(2,0,0)[7]             : OOB=10629153985.559, Time=0.43 sec
 ARIMA(2,1,0)(2,0,1)[7]             : OOB=1620921487.928, Time=0.91 sec
 ARIMA(2,1,0)(2,0,2)[7]             : OOB=1617426288.392, Time=1.15 sec
 ARIMA(2,1,0)(2,0,3)[7]             : OOB=1892779452.340, Time=2.27 sec
 ARIMA(2,1,0)(2,0,4)[7]             : OOB=1641527418.466, Time=2.84 sec
 ARIMA(2,1,0)(3,0,0)[7]             : OOB=12244455118.084, Time=0.89 sec
 ARIMA(2,1,0)(3,0,1)[7]             : OOB=1579380693.081, Time=2.21 sec
 ARIMA(2,1,0)(3,0,2)[7]             : OOB=1680298660.258, Time=1.23 sec
 ARIMA(2,1,0)(3,0,3)[7]             : OOB=1821262833.782, Time=2.51 sec
 ARIMA(2,1,0)(4,0,0)[7]             : OOB=inf, Time=2.63 sec
 ARIMA(2,1,0)(4,0,1)[7]             : OOB=1867098040.222, Time=3.23 sec
 ARIMA(2,1,0)(4,0,2)[7]             : OOB=1658545625.074, Time=3.48 sec
 ARIMA(2,1,0)(5,0,0)[7]             : OOB=inf, Time=3.05 sec
 ARIMA(2,1,0)(5,0,1)[7]             : OOB=2028378236.820, Time=5.69 sec
 ARIMA(2,1,1)(0,0,0)[7]             : OOB=10415340580.478, Time=0.73 sec
 ARIMA(2,1,1)(0,0,1)[7]             : OOB=5516810699.962, Time=0.94 sec
 ARIMA(2,1,1)(0,0,2)[7]             : OOB=3323233436.971, Time=0.73 sec
 ARIMA(2,1,1)(0,0,3)[7]             : OOB=1521665941.283, Time=1.37 sec
 ARIMA(2,1,1)(0,0,4)[7]             : OOB=1311511665.737, Time=2.56 sec
 ARIMA(2,1,1)(0,0,5)[7]             : OOB=2033618362.894, Time=3.78 sec
 ARIMA(2,1,1)(1,0,0)[7]             : OOB=4190228429.309, Time=0.44 sec
 ARIMA(2,1,1)(1,0,1)[7]             : OOB=9338391504.462, Time=1.05 sec
 ARIMA(2,1,1)(1,0,2)[7]             : OOB=3026539738.562, Time=0.93 sec
 ARIMA(2,1,1)(1,0,3)[7]             : OOB=1654698562.873, Time=1.55 sec
 ARIMA(2,1,1)(1,0,4)[7]             : OOB=1383182706.951, Time=3.27 sec
 ARIMA(2,1,1)(2,0,0)[7]             : OOB=9166837621.684, Time=0.78 sec
 ARIMA(2,1,1)(2,0,1)[7]             : OOB=1396942924.004, Time=1.21 sec
 ARIMA(2,1,1)(2,0,2)[7]             : OOB=1474009859.397, Time=1.37 sec
 ARIMA(2,1,1)(2,0,3)[7]             : OOB=1756992216.283, Time=2.81 sec
 ARIMA(2,1,1)(3,0,0)[7]             : OOB=12107637537.063, Time=1.22 sec
 ARIMA(2,1,1)(3,0,1)[7]             : OOB=1384000653.377, Time=1.69 sec
 ARIMA(2,1,1)(3,0,2)[7]             : OOB=1430190254.723, Time=3.59 sec
 ARIMA(2,1,1)(4,0,0)[7]             : OOB=11508432520.549, Time=2.18 sec
 ARIMA(2,1,1)(4,0,1)[7]             : OOB=1819376894.097, Time=4.03 sec
 ARIMA(2,1,1)(5,0,0)[7]             : OOB=9919274078.840, Time=3.47 sec
 ARIMA(2,1,2)(0,0,0)[7]             : OOB=4468194877.744, Time=0.41 sec
 ARIMA(2,1,2)(0,0,1)[7]             : OOB=2670403117.845, Time=1.35 sec
 ARIMA(2,1,2)(0,0,2)[7]             : OOB=2271553679.648, Time=1.62 sec
 ARIMA(2,1,2)(0,0,3)[7]             : OOB=957060700.079, Time=2.45 sec
 ARIMA(2,1,2)(0,0,4)[7]             : OOB=1539013070.804, Time=3.80 sec
 ARIMA(2,1,2)(1,0,0)[7]             : OOB=5266949321.393, Time=1.27 sec
 ARIMA(2,1,2)(1,0,1)[7]             : OOB=10379934760.724, Time=1.29 sec
 ARIMA(2,1,2)(1,0,2)[7]             : OOB=2299944512.381, Time=1.80 sec
 ARIMA(2,1,2)(1,0,3)[7]             : OOB=1583243107.838, Time=3.80 sec
 ARIMA(2,1,2)(2,0,0)[7]             : OOB=9441227978.058, Time=1.43 sec
 ARIMA(2,1,2)(2,0,1)[7]             : OOB=1273687689.119, Time=1.94 sec
 ARIMA(2,1,2)(2,0,2)[7]             : OOB=1267615680.365, Time=2.24 sec
 ARIMA(2,1,2)(3,0,0)[7]             : OOB=12786303370.281, Time=2.56 sec
 ARIMA(2,1,2)(3,0,1)[7]             : OOB=1265525970.674, Time=2.99 sec
 ARIMA(2,1,2)(4,0,0)[7]             : OOB=12261428803.093, Time=4.03 sec
 ARIMA(2,1,3)(0,0,0)[7]             : OOB=10631672116.542, Time=0.35 sec
 ARIMA(2,1,3)(0,0,1)[7]             : OOB=4434013098.961, Time=2.36 sec
 ARIMA(2,1,3)(0,0,2)[7]             : OOB=3131862215.282, Time=2.90 sec
 ARIMA(2,1,3)(0,0,3)[7]             : OOB=1702753985.743, Time=4.96 sec
 ARIMA(2,1,3)(1,0,0)[7]             : OOB=1176110727.686, Time=1.19 sec
 ARIMA(2,1,3)(1,0,1)[7]             : OOB=inf, Time=2.45 sec
 ARIMA(2,1,3)(1,0,2)[7]             : OOB=2470521126.223, Time=2.16 sec
 ARIMA(2,1,3)(2,0,0)[7]             : OOB=inf, Time=2.83 sec
 ARIMA(2,1,3)(2,0,1)[7]             : OOB=1413422724.963, Time=2.01 sec
 ARIMA(2,1,3)(3,0,0)[7]             : OOB=13797917996.427, Time=2.83 sec
 ARIMA(2,1,4)(0,0,0)[7]             : OOB=7459895802.660, Time=0.83 sec
 ARIMA(2,1,4)(0,0,1)[7]             : OOB=4192200752.589, Time=2.17 sec
 ARIMA(2,1,4)(0,0,2)[7]             : OOB=3020294527.963, Time=3.15 sec
 ARIMA(2,1,4)(1,0,0)[7]             : OOB=1302374508.127, Time=1.42 sec
 ARIMA(2,1,4)(1,0,1)[7]             : OOB=7466689867.965, Time=1.28 sec
 ARIMA(2,1,4)(2,0,0)[7]             : OOB=707904819.113, Time=3.65 sec
 ARIMA(2,1,5)(0,0,0)[7]             : OOB=5260710697.209, Time=1.10 sec
 ARIMA(2,1,5)(0,0,1)[7]             : OOB=3717889915.705, Time=2.28 sec
 ARIMA(2,1,5)(1,0,0)[7]             : OOB=1324149999.179, Time=2.98 sec
 ARIMA(3,1,0)(0,0,0)[7]             : OOB=4322448813.065, Time=0.23 sec
 ARIMA(3,1,0)(0,0,1)[7]             : OOB=2006120511.550, Time=0.38 sec
 ARIMA(3,1,0)(0,0,2)[7]             : OOB=1705777022.107, Time=0.46 sec
 ARIMA(3,1,0)(0,0,3)[7]             : OOB=2530583458.535, Time=0.96 sec
 ARIMA(3,1,0)(0,0,4)[7]             : OOB=3394449848.456, Time=1.59 sec
 ARIMA(3,1,0)(0,0,5)[7]             : OOB=4370697158.820, Time=3.04 sec
 ARIMA(3,1,0)(1,0,0)[7]             : OOB=9150606130.340, Time=1.08 sec
 ARIMA(3,1,0)(1,0,1)[7]             : OOB=9959650674.149, Time=1.62 sec
 ARIMA(3,1,0)(1,0,2)[7]             : OOB=2994394436.129, Time=0.71 sec
 ARIMA(3,1,0)(1,0,3)[7]             : OOB=1919067256.767, Time=1.24 sec
 ARIMA(3,1,0)(1,0,4)[7]             : OOB=1643094147.234, Time=2.32 sec
 ARIMA(3,1,0)(2,0,0)[7]             : OOB=11264729119.512, Time=0.91 sec
 ARIMA(3,1,0)(2,0,1)[7]             : OOB=1627547040.628, Time=1.11 sec
 ARIMA(3,1,0)(2,0,2)[7]             : OOB=1644934353.128, Time=1.21 sec
 ARIMA(3,1,0)(2,0,3)[7]             : OOB=2932771944.504, Time=1.42 sec
 ARIMA(3,1,0)(3,0,0)[7]             : OOB=13632309407.748, Time=1.45 sec
 ARIMA(3,1,0)(3,0,1)[7]             : OOB=1620789820.362, Time=1.82 sec
 ARIMA(3,1,0)(3,0,2)[7]             : OOB=1696483439.103, Time=2.25 sec
 ARIMA(3,1,0)(4,0,0)[7]             : OOB=inf, Time=2.41 sec
 ARIMA(3,1,0)(4,0,1)[7]             : OOB=2111544677.085, Time=4.18 sec
 ARIMA(3,1,0)(5,0,0)[7]             : OOB=inf, Time=3.69 sec
 ARIMA(3,1,1)(0,0,0)[7]             : OOB=8966652238.541, Time=0.68 sec
 ARIMA(3,1,1)(0,0,1)[7]             : OOB=4261088976.425, Time=0.82 sec
 ARIMA(3,1,1)(0,0,2)[7]             : OOB=2624140357.391, Time=0.79 sec
 ARIMA(3,1,1)(0,0,3)[7]             : OOB=1267616691.330, Time=1.17 sec
 ARIMA(3,1,1)(0,0,4)[7]             : OOB=1348346834.933, Time=2.00 sec
 ARIMA(3,1,1)(1,0,0)[7]             : OOB=4895760486.292, Time=0.88 sec
 ARIMA(3,1,1)(1,0,1)[7]             : OOB=9472165443.526, Time=1.24 sec
 ARIMA(3,1,1)(1,0,2)[7]             : OOB=2750435052.190, Time=1.01 sec
 ARIMA(3,1,1)(1,0,3)[7]             : OOB=1580381175.554, Time=2.14 sec
 ARIMA(3,1,1)(2,0,0)[7]             : OOB=9062212363.546, Time=0.75 sec
 ARIMA(3,1,1)(2,0,1)[7]             : OOB=1330430653.912, Time=1.23 sec
 ARIMA(3,1,1)(2,0,2)[7]             : OOB=1402533340.236, Time=2.07 sec
 ARIMA(3,1,1)(3,0,0)[7]             : OOB=12205699648.362, Time=1.42 sec
 ARIMA(3,1,1)(3,0,1)[7]             : OOB=1333949962.468, Time=2.34 sec
 ARIMA(3,1,1)(4,0,0)[7]             : OOB=11745613655.192, Time=3.85 sec
 ARIMA(3,1,2)(0,0,0)[7]             : OOB=5880652134.554, Time=0.57 sec
 ARIMA(3,1,2)(0,0,1)[7]             : OOB=3376041297.760, Time=2.35 sec
 ARIMA(3,1,2)(0,0,2)[7]             : OOB=2304442252.491, Time=2.11 sec
 ARIMA(3,1,2)(0,0,3)[7]             : OOB=1227129561.388, Time=4.43 sec
 ARIMA(3,1,2)(1,0,0)[7]             : OOB=5327649301.675, Time=1.40 sec
 ARIMA(3,1,2)(1,0,1)[7]             : OOB=7934771846.493, Time=2.55 sec
 ARIMA(3,1,2)(1,0,2)[7]             : OOB=1237059328.072, Time=1.62 sec
 ARIMA(3,1,2)(2,0,0)[7]             : OOB=9554174988.728, Time=1.79 sec
 ARIMA(3,1,2)(2,0,1)[7]             : OOB=999726684.251, Time=1.81 sec
 ARIMA(3,1,2)(3,0,0)[7]             : OOB=11934415597.170, Time=2.45 sec
 ARIMA(3,1,3)(0,0,0)[7]             : OOB=7682201374.063, Time=1.56 sec
 ARIMA(3,1,3)(0,0,1)[7]             : OOB=4318672324.816, Time=3.33 sec
 ARIMA(3,1,3)(0,0,2)[7]             : OOB=3096326019.669, Time=5.92 sec
 ARIMA(3,1,3)(1,0,0)[7]             : OOB=534872731.234, Time=1.64 sec
 ARIMA(3,1,3)(1,0,1)[7]             : OOB=9032424168.819, Time=1.44 sec
 ARIMA(3,1,3)(2,0,0)[7]             : OOB=9955001085.845, Time=1.26 sec
 ARIMA(3,1,4)(0,0,0)[7]             : OOB=7695726326.674, Time=0.75 sec
 ARIMA(3,1,4)(0,0,1)[7]             : OOB=3861382758.804, Time=2.36 sec
 ARIMA(3,1,4)(1,0,0)[7]             : OOB=979838770.766, Time=1.50 sec
 ARIMA(3,1,5)(0,0,0)[7]             : OOB=2416455693.283, Time=2.53 sec
 ARIMA(4,1,0)(0,0,0)[7]             : OOB=6214462214.913, Time=0.21 sec
 ARIMA(4,1,0)(0,0,1)[7]             : OOB=2629306286.150, Time=0.61 sec
 ARIMA(4,1,0)(0,0,2)[7]             : OOB=1680486397.286, Time=0.64 sec
 ARIMA(4,1,0)(0,0,3)[7]             : OOB=1476623954.237, Time=1.12 sec
 ARIMA(4,1,0)(0,0,4)[7]             : OOB=2076487127.884, Time=2.51 sec
 ARIMA(4,1,0)(1,0,0)[7]             : OOB=8326137388.806, Time=0.89 sec
 ARIMA(4,1,0)(1,0,1)[7]             : OOB=9971038286.719, Time=1.14 sec
 ARIMA(4,1,0)(1,0,2)[7]             : OOB=2874786051.068, Time=0.88 sec
 ARIMA(4,1,0)(1,0,3)[7]             : OOB=1786608016.375, Time=1.73 sec
 ARIMA(4,1,0)(2,0,0)[7]             : OOB=10819812470.228, Time=0.76 sec
 ARIMA(4,1,0)(2,0,1)[7]             : OOB=1535830721.078, Time=1.19 sec
 ARIMA(4,1,0)(2,0,2)[7]             : OOB=1564037641.059, Time=1.79 sec
 ARIMA(4,1,0)(3,0,0)[7]             : OOB=13629614452.686, Time=1.50 sec
 ARIMA(4,1,0)(3,0,1)[7]             : OOB=1530220130.370, Time=2.43 sec
 ARIMA(4,1,0)(4,0,0)[7]             : OOB=inf, Time=2.90 sec
 ARIMA(4,1,1)(0,0,0)[7]             : OOB=8744444045.507, Time=0.62 sec
 ARIMA(4,1,1)(0,0,1)[7]             : OOB=4232649948.752, Time=0.49 sec
 ARIMA(4,1,1)(0,0,2)[7]             : OOB=2640418133.153, Time=0.64 sec
 ARIMA(4,1,1)(0,0,3)[7]             : OOB=1215942755.689, Time=1.17 sec
 ARIMA(4,1,1)(1,0,0)[7]             : OOB=3227945687.739, Time=0.80 sec
 ARIMA(4,1,1)(1,0,1)[7]             : OOB=9646264694.342, Time=1.12 sec
 ARIMA(4,1,1)(1,0,2)[7]             : OOB=2721749798.328, Time=1.32 sec
 ARIMA(4,1,1)(2,0,0)[7]             : OOB=9323576156.224, Time=1.08 sec
 ARIMA(4,1,1)(2,0,1)[7]             : OOB=1248089611.191, Time=1.61 sec
 ARIMA(4,1,1)(3,0,0)[7]             : OOB=12613010740.386, Time=2.09 sec
 ARIMA(4,1,2)(0,0,0)[7]             : OOB=7467526169.781, Time=0.92 sec
 ARIMA(4,1,2)(0,0,1)[7]             : OOB=4137172767.244, Time=1.61 sec
 ARIMA(4,1,2)(0,0,2)[7]             : OOB=2912607131.017, Time=2.46 sec
 ARIMA(4,1,2)(1,0,0)[7]             : OOB=1299954317.426, Time=1.24 sec
 ARIMA(4,1,2)(1,0,1)[7]             : OOB=8035307332.437, Time=1.57 sec
 ARIMA(4,1,2)(2,0,0)[7]             : OOB=753945822.061, Time=3.95 sec
 ARIMA(4,1,3)(0,0,0)[7]             : OOB=5894035045.765, Time=1.35 sec
 ARIMA(4,1,3)(0,0,1)[7]             : OOB=3224303307.236, Time=3.67 sec
 ARIMA(4,1,3)(1,0,0)[7]             : OOB=inf, Time=nan sec
 ARIMA(4,1,4)(0,0,0)[7]             : OOB=7618546285.978, Time=1.18 sec
 ARIMA(5,1,0)(0,0,0)[7]             : OOB=8604643113.648, Time=0.26 sec
 ARIMA(5,1,0)(0,0,1)[7]             : OOB=4079282878.525, Time=0.96 sec
 ARIMA(5,1,0)(0,0,2)[7]             : OOB=2675386811.584, Time=0.64 sec
 ARIMA(5,1,0)(0,0,3)[7]             : OOB=1208671450.536, Time=1.25 sec
 ARIMA(5,1,0)(1,0,0)[7]             : OOB=2721089633.050, Time=1.58 sec
 ARIMA(5,1,0)(1,0,1)[7]             : OOB=9335079098.064, Time=0.86 sec
 ARIMA(5,1,0)(1,0,2)[7]             : OOB=2118505021.799, Time=1.01 sec
 ARIMA(5,1,0)(2,0,0)[7]             : OOB=7949539256.572, Time=1.18 sec
 ARIMA(5,1,0)(2,0,1)[7]             : OOB=1140631695.294, Time=2.10 sec
 ARIMA(5,1,0)(3,0,0)[7]             : OOB=11881455616.213, Time=1.87 sec
 ARIMA(5,1,1)(0,0,0)[7]             : OOB=8958118472.915, Time=0.39 sec
 ARIMA(5,1,1)(0,0,1)[7]             : OOB=4239551878.865, Time=1.03 sec
 ARIMA(5,1,1)(0,0,2)[7]             : OOB=2775079153.362, Time=1.07 sec
 ARIMA(5,1,1)(1,0,0)[7]             : OOB=5862896625.965, Time=1.64 sec
 ARIMA(5,1,1)(1,0,1)[7]             : OOB=9291406149.196, Time=1.40 sec
 ARIMA(5,1,1)(2,0,0)[7]             : OOB=8860128303.179, Time=1.79 sec
 ARIMA(5,1,2)(0,0,0)[7]             : OOB=7598447777.567, Time=0.89 sec
 ARIMA(5,1,2)(0,0,1)[7]             : OOB=4102333189.554, Time=1.74 sec
 ARIMA(5,1,2)(1,0,0)[7]             : OOB=10120481383.037, Time=1.74 sec
 ARIMA(5,1,3)(0,0,0)[7]             : OOB=7562332751.858, Time=1.35 sec

Best model:  ARIMA(3,1,3)(1,0,0)[7]          
Total fit time: 1158.202 seconds

ARIMA(maxiter=200, order=(3, 1, 3), out_of_sample_size=30, scoring_args={},
      seasonal_order=(1, 0, 0, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((3,1,3),(1,0,0,7)), aic_bic=False).rmse()

21544.510408173162


#Might take some time to finish

auto_arima(train_df.new_cases.append(val_df.new_cases), start_p=0, d=1, start_q=0, max_p=5, max_d=2, max_q=5,
           start_P=0, D=1, start_Q=0, max_P=5, max_D=1, max_Q=5, seasonal=True, m=7, maxiter=200, trace=True,
           out_of_sample_size=30, information_criterion='oob', with_intercept=False, stepwise=False, max_order=8)

 ARIMA(0,1,0)(0,1,0)[7]             : OOB=15416368714.367, Time=0.18 sec
 ARIMA(0,1,0)(0,1,1)[7]             : OOB=3358794967.374, Time=0.16 sec
 ARIMA(0,1,0)(0,1,2)[7]             : OOB=2063013999.117, Time=0.50 sec
 ARIMA(0,1,0)(0,1,3)[7]             : OOB=1946207990.052, Time=0.90 sec
 ARIMA(0,1,0)(0,1,4)[7]             : OOB=1697178652.985, Time=1.79 sec
 ARIMA(0,1,0)(0,1,5)[7]             : OOB=1558944781.557, Time=2.21 sec
 ARIMA(0,1,0)(1,1,0)[7]             : OOB=16599438472.908, Time=0.20 sec
 ARIMA(0,1,0)(1,1,1)[7]             : OOB=1911003079.620, Time=0.26 sec
 ARIMA(0,1,0)(1,1,2)[7]             : OOB=1593889249.736, Time=0.80 sec
 ARIMA(0,1,0)(1,1,3)[7]             : OOB=1597603678.054, Time=1.40 sec
 ARIMA(0,1,0)(1,1,4)[7]             : OOB=1867058241.079, Time=2.56 sec
 ARIMA(0,1,0)(1,1,5)[7]             : OOB=1818458675.844, Time=3.73 sec
 ARIMA(0,1,0)(2,1,0)[7]             : OOB=16703479723.887, Time=0.46 sec
 ARIMA(0,1,0)(2,1,1)[7]             : OOB=1655346689.754, Time=0.68 sec
 ARIMA(0,1,0)(2,1,2)[7]             : OOB=1918399302.549, Time=0.63 sec
 ARIMA(0,1,0)(2,1,3)[7]             : OOB=1545226251.608, Time=2.66 sec
 ARIMA(0,1,0)(2,1,4)[7]             : OOB=1593323245.098, Time=4.12 sec
 ARIMA(0,1,0)(2,1,5)[7]             : OOB=1612712400.351, Time=5.96 sec
 ARIMA(0,1,0)(3,1,0)[7]             : OOB=14970153898.810, Time=1.44 sec
 ARIMA(0,1,0)(3,1,1)[7]             : OOB=1567389917.454, Time=1.18 sec
 ARIMA(0,1,0)(3,1,2)[7]             : OOB=1540756356.921, Time=1.95 sec
 ARIMA(0,1,0)(3,1,3)[7]             : OOB=inf, Time=3.72 sec
 ARIMA(0,1,0)(3,1,4)[7]             : OOB=inf, Time=8.06 sec
 ARIMA(0,1,0)(3,1,5)[7]             : OOB=inf, Time=14.92 sec
 ARIMA(0,1,0)(4,1,0)[7]             : OOB=10917097104.561, Time=1.02 sec
 ARIMA(0,1,0)(4,1,1)[7]             : OOB=1625105319.802, Time=3.30 sec
 ARIMA(0,1,0)(4,1,2)[7]             : OOB=1669746856.610, Time=2.94 sec
 ARIMA(0,1,0)(4,1,3)[7]             : OOB=inf, Time=6.47 sec
 ARIMA(0,1,0)(4,1,4)[7]             : OOB=inf, Time=13.15 sec
 ARIMA(0,1,0)(5,1,0)[7]             : OOB=9475260224.515, Time=2.48 sec
 ARIMA(0,1,0)(5,1,1)[7]             : OOB=1677290792.636, Time=4.26 sec
 ARIMA(0,1,0)(5,1,2)[7]             : OOB=1559972679.625, Time=4.84 sec
 ARIMA(0,1,0)(5,1,3)[7]             : OOB=1519062595.324, Time=6.93 sec
 ARIMA(0,1,1)(0,1,0)[7]             : OOB=22555891059.215, Time=0.10 sec
 ARIMA(0,1,1)(0,1,1)[7]             : OOB=10456550693.569, Time=1.80 sec
 ARIMA(0,1,1)(0,1,2)[7]             : OOB=2278506099.589, Time=2.08 sec
 ARIMA(0,1,1)(0,1,3)[7]             : OOB=1661370308.723, Time=1.46 sec
 ARIMA(0,1,1)(0,1,4)[7]             : OOB=1528017041.130, Time=1.63 sec
 ARIMA(0,1,1)(0,1,5)[7]             : OOB=1584481574.306, Time=3.58 sec
 ARIMA(0,1,1)(1,1,0)[7]             : OOB=20295837051.377, Time=0.20 sec
 ARIMA(0,1,1)(1,1,1)[7]             : OOB=1558180381.616, Time=0.37 sec
 ARIMA(0,1,1)(1,1,2)[7]             : OOB=1604827627.095, Time=0.85 sec
 ARIMA(0,1,1)(1,1,3)[7]             : OOB=1794409838.581, Time=41.93 sec
 ARIMA(0,1,1)(1,1,4)[7]             : OOB=1557511580.208, Time=3.27 sec
 ARIMA(0,1,1)(1,1,5)[7]             : OOB=1870766375.833, Time=4.90 sec
 ARIMA(0,1,1)(2,1,0)[7]             : OOB=21629859597.917, Time=0.41 sec
 ARIMA(0,1,1)(2,1,1)[7]             : OOB=1556228839.414, Time=0.75 sec
 ARIMA(0,1,1)(2,1,2)[7]             : OOB=1568382379.992, Time=1.02 sec
 ARIMA(0,1,1)(2,1,3)[7]             : OOB=1779554011.936, Time=1.62 sec
 ARIMA(0,1,1)(2,1,4)[7]             : OOB=1775935784.952, Time=4.73 sec
 ARIMA(0,1,1)(2,1,5)[7]             : OOB=1781898229.171, Time=4.83 sec
 ARIMA(0,1,1)(3,1,0)[7]             : OOB=18382330909.786, Time=0.78 sec
 ARIMA(0,1,1)(3,1,1)[7]             : OOB=1842119099.850, Time=1.54 sec
 ARIMA(0,1,1)(3,1,2)[7]             : OOB=1636033463.728, Time=1.84 sec
 ARIMA(0,1,1)(3,1,3)[7]             : OOB=1558507672.904, Time=2.36 sec
 ARIMA(0,1,1)(3,1,4)[7]             : OOB=inf, Time=10.35 sec
 ARIMA(0,1,1)(4,1,0)[7]             : OOB=14637406759.085, Time=1.87 sec
 ARIMA(0,1,1)(4,1,1)[7]             : OOB=1882674625.676, Time=2.98 sec
 ARIMA(0,1,1)(4,1,2)[7]             : OOB=1735589341.662, Time=3.84 sec
 ARIMA(0,1,1)(4,1,3)[7]             : OOB=1663400520.196, Time=4.31 sec
 ARIMA(0,1,1)(5,1,0)[7]             : OOB=12889967799.259, Time=2.59 sec
 ARIMA(0,1,1)(5,1,1)[7]             : OOB=1727854335.878, Time=5.08 sec
 ARIMA(0,1,1)(5,1,2)[7]             : OOB=2101049606.297, Time=5.74 sec
 ARIMA(0,1,2)(0,1,0)[7]             : OOB=22121058293.314, Time=0.35 sec
 ARIMA(0,1,2)(0,1,1)[7]             : OOB=10449052791.306, Time=0.30 sec
 ARIMA(0,1,2)(0,1,2)[7]             : OOB=2261760783.428, Time=0.83 sec
 ARIMA(0,1,2)(0,1,3)[7]             : OOB=1644188803.241, Time=1.55 sec
 ARIMA(0,1,2)(0,1,4)[7]             : OOB=1507039143.352, Time=2.94 sec
 ARIMA(0,1,2)(0,1,5)[7]             : OOB=1559843060.343, Time=3.60 sec
 ARIMA(0,1,2)(1,1,0)[7]             : OOB=20419485094.016, Time=0.32 sec
 ARIMA(0,1,2)(1,1,1)[7]             : OOB=1542951525.476, Time=0.54 sec
 ARIMA(0,1,2)(1,1,2)[7]             : OOB=1592035550.904, Time=1.36 sec
 ARIMA(0,1,2)(1,1,3)[7]             : OOB=1769726530.394, Time=2.77 sec
 ARIMA(0,1,2)(1,1,4)[7]             : OOB=1540883899.620, Time=4.73 sec
 ARIMA(0,1,2)(1,1,5)[7]             : OOB=1837748419.539, Time=7.04 sec
 ARIMA(0,1,2)(2,1,0)[7]             : OOB=21731874264.782, Time=0.51 sec
 ARIMA(0,1,2)(2,1,1)[7]             : OOB=1536535774.280, Time=1.07 sec
 ARIMA(0,1,2)(2,1,2)[7]             : OOB=1552237428.723, Time=1.53 sec
 ARIMA(0,1,2)(2,1,3)[7]             : OOB=1768454994.626, Time=2.46 sec
 ARIMA(0,1,2)(2,1,4)[7]             : OOB=1752246365.053, Time=5.71 sec
 ARIMA(0,1,2)(3,1,0)[7]             : OOB=18426686789.816, Time=0.96 sec
 ARIMA(0,1,2)(3,1,1)[7]             : OOB=1822055811.996, Time=1.78 sec
 ARIMA(0,1,2)(3,1,2)[7]             : OOB=1616237391.760, Time=2.11 sec
 ARIMA(0,1,2)(3,1,3)[7]             : OOB=1531808993.449, Time=3.86 sec
 ARIMA(0,1,2)(4,1,0)[7]             : OOB=14633026027.053, Time=1.80 sec
 ARIMA(0,1,2)(4,1,1)[7]             : OOB=1854703734.334, Time=7.91 sec
 ARIMA(0,1,2)(4,1,2)[7]             : OOB=1715165497.805, Time=10.72 sec
 ARIMA(0,1,2)(5,1,0)[7]             : OOB=12894803069.627, Time=3.89 sec
 ARIMA(0,1,2)(5,1,1)[7]             : OOB=1679717012.541, Time=7.16 sec
 ARIMA(0,1,3)(0,1,0)[7]             : OOB=22949314525.666, Time=0.91 sec
 ARIMA(0,1,3)(0,1,1)[7]             : OOB=10707751764.272, Time=4.46 sec
 ARIMA(0,1,3)(0,1,2)[7]             : OOB=2127257892.283, Time=2.54 sec
 ARIMA(0,1,3)(0,1,3)[7]             : OOB=1591257504.657, Time=2.69 sec
 ARIMA(0,1,3)(0,1,4)[7]             : OOB=1430494518.449, Time=3.52 sec
 ARIMA(0,1,3)(0,1,5)[7]             : OOB=1462728733.244, Time=5.40 sec
 ARIMA(0,1,3)(1,1,0)[7]             : OOB=21014108690.376, Time=0.76 sec
 ARIMA(0,1,3)(1,1,1)[7]             : OOB=1471759384.828, Time=0.72 sec
 ARIMA(0,1,3)(1,1,2)[7]             : OOB=1506505813.419, Time=1.43 sec
 ARIMA(0,1,3)(1,1,3)[7]             : OOB=1669505731.013, Time=2.99 sec
 ARIMA(0,1,3)(1,1,4)[7]             : OOB=1457258341.206, Time=3.52 sec
 ARIMA(0,1,3)(2,1,0)[7]             : OOB=22104477133.955, Time=1.15 sec
 ARIMA(0,1,3)(2,1,1)[7]             : OOB=1464738560.612, Time=1.02 sec
 ARIMA(0,1,3)(2,1,2)[7]             : OOB=1476786173.373, Time=3.08 sec
 ARIMA(0,1,3)(2,1,3)[7]             : OOB=1659036336.807, Time=4.18 sec
 ARIMA(0,1,3)(3,1,0)[7]             : OOB=18756669188.762, Time=1.77 sec
 ARIMA(0,1,3)(3,1,1)[7]             : OOB=1734899695.048, Time=2.54 sec
 ARIMA(0,1,3)(3,1,2)[7]             : OOB=1532947215.083, Time=3.88 sec
 ARIMA(0,1,3)(4,1,0)[7]             : OOB=14840864164.142, Time=2.56 sec
 ARIMA(0,1,3)(4,1,1)[7]             : OOB=1694613025.670, Time=5.04 sec
 ARIMA(0,1,3)(5,1,0)[7]             : OOB=12931554832.871, Time=3.15 sec
 ARIMA(0,1,4)(0,1,0)[7]             : OOB=22959492980.149, Time=0.70 sec
 ARIMA(0,1,4)(0,1,1)[7]             : OOB=10584458907.608, Time=0.58 sec
 ARIMA(0,1,4)(0,1,2)[7]             : OOB=2170528023.256, Time=2.15 sec
 ARIMA(0,1,4)(0,1,3)[7]             : OOB=1621634658.712, Time=2.33 sec
 ARIMA(0,1,4)(0,1,4)[7]             : OOB=1482013936.286, Time=4.29 sec
 ARIMA(0,1,4)(1,1,0)[7]             : OOB=21061970681.596, Time=2.64 sec
 ARIMA(0,1,4)(1,1,1)[7]             : OOB=1510576743.104, Time=1.44 sec
 ARIMA(0,1,4)(1,1,2)[7]             : OOB=1557215968.002, Time=58.10 sec
 ARIMA(0,1,4)(1,1,3)[7]             : OOB=1756625142.790, Time=7.94 sec
 ARIMA(0,1,4)(2,1,0)[7]             : OOB=22517804817.763, Time=4.72 sec
 ARIMA(0,1,4)(2,1,1)[7]             : OOB=1513671472.960, Time=5.19 sec
 ARIMA(0,1,4)(2,1,2)[7]             : OOB=1522205035.433, Time=4.22 sec
 ARIMA(0,1,4)(3,1,0)[7]             : OOB=19177297802.944, Time=3.06 sec
 ARIMA(0,1,4)(3,1,1)[7]             : OOB=1845247012.746, Time=3.73 sec
 ARIMA(0,1,4)(4,1,0)[7]             : OOB=15219157832.002, Time=7.39 sec
 ARIMA(0,1,5)(0,1,0)[7]             : OOB=23590612412.038, Time=3.76 sec
 ARIMA(0,1,5)(0,1,1)[7]             : OOB=10602771322.112, Time=2.13 sec
 ARIMA(0,1,5)(0,1,2)[7]             : OOB=2174722853.431, Time=3.14 sec
 ARIMA(0,1,5)(0,1,3)[7]             : OOB=1623400057.825, Time=4.06 sec
 ARIMA(0,1,5)(1,1,0)[7]             : OOB=21413373254.516, Time=1.39 sec
 ARIMA(0,1,5)(1,1,1)[7]             : OOB=1512056962.833, Time=1.30 sec
 ARIMA(0,1,5)(1,1,2)[7]             : OOB=1557637435.412, Time=3.06 sec
 ARIMA(0,1,5)(2,1,0)[7]             : OOB=22321534552.039, Time=1.62 sec
 ARIMA(0,1,5)(2,1,1)[7]             : OOB=1514424406.287, Time=1.36 sec
 ARIMA(0,1,5)(3,1,0)[7]             : OOB=18953770618.137, Time=1.27 sec
 ARIMA(1,1,0)(0,1,0)[7]             : OOB=17515050376.610, Time=1.18 sec
 ARIMA(1,1,0)(0,1,1)[7]             : OOB=8140180059.525, Time=0.90 sec
 ARIMA(1,1,0)(0,1,2)[7]             : OOB=1952771954.390, Time=2.36 sec
 ARIMA(1,1,0)(0,1,3)[7]             : OOB=1649783921.245, Time=1.85 sec
 ARIMA(1,1,0)(0,1,4)[7]             : OOB=1502252478.702, Time=2.82 sec
 ARIMA(1,1,0)(0,1,5)[7]             : OOB=1493125310.182, Time=4.18 sec
 ARIMA(1,1,0)(1,1,0)[7]             : OOB=17276749347.203, Time=0.53 sec
 ARIMA(1,1,0)(1,1,1)[7]             : OOB=1537705393.372, Time=0.75 sec
 ARIMA(1,1,0)(1,1,2)[7]             : OOB=1512583823.170, Time=1.91 sec
 ARIMA(1,1,0)(1,1,3)[7]             : OOB=1588849446.861, Time=1.99 sec
 ARIMA(1,1,0)(1,1,4)[7]             : OOB=1538366361.099, Time=3.54 sec
 ARIMA(1,1,0)(1,1,5)[7]             : OOB=1509762039.440, Time=6.99 sec
 ARIMA(1,1,0)(2,1,0)[7]             : OOB=17489815051.118, Time=1.97 sec
 ARIMA(1,1,0)(2,1,1)[7]             : OOB=1515118308.985, Time=1.70 sec
 ARIMA(1,1,0)(2,1,2)[7]             : OOB=1546050781.414, Time=2.49 sec
 ARIMA(1,1,0)(2,1,3)[7]             : OOB=1539692473.681, Time=4.05 sec
 ARIMA(1,1,0)(2,1,4)[7]             : OOB=1527211342.950, Time=4.87 sec
 ARIMA(1,1,0)(2,1,5)[7]             : OOB=1529907264.774, Time=9.75 sec
 ARIMA(1,1,0)(3,1,0)[7]             : OOB=15166387887.887, Time=1.87 sec
 ARIMA(1,1,0)(3,1,1)[7]             : OOB=1588633611.045, Time=1.79 sec
 ARIMA(1,1,0)(3,1,2)[7]             : OOB=1500639029.672, Time=12.01 sec
 ARIMA(1,1,0)(3,1,3)[7]             : OOB=inf, Time=8.11 sec
 ARIMA(1,1,0)(3,1,4)[7]             : OOB=inf, Time=13.09 sec
 ARIMA(1,1,0)(4,1,0)[7]             : OOB=11534032848.820, Time=4.82 sec
 ARIMA(1,1,0)(4,1,1)[7]             : OOB=1574002094.899, Time=6.11 sec
 ARIMA(1,1,0)(4,1,2)[7]             : OOB=1500060566.949, Time=5.59 sec
 ARIMA(1,1,0)(4,1,3)[7]             : OOB=inf, Time=13.20 sec
 ARIMA(1,1,0)(5,1,0)[7]             : OOB=10120549564.443, Time=7.24 sec
 ARIMA(1,1,0)(5,1,1)[7]             : OOB=1578833770.380, Time=6.03 sec
 ARIMA(1,1,0)(5,1,2)[7]             : OOB=1721233409.962, Time=6.81 sec
 ARIMA(1,1,1)(0,1,0)[7]             : OOB=21951757056.450, Time=1.99 sec
 ARIMA(1,1,1)(0,1,1)[7]             : OOB=10444491695.333, Time=3.47 sec
 ARIMA(1,1,1)(0,1,2)[7]             : OOB=2240745193.379, Time=2.75 sec
 ARIMA(1,1,1)(0,1,3)[7]             : OOB=1630500972.789, Time=2.23 sec
 ARIMA(1,1,1)(0,1,4)[7]             : OOB=1491753026.221, Time=7.99 sec
 ARIMA(1,1,1)(0,1,5)[7]             : OOB=1540952905.182, Time=12.43 sec
 ARIMA(1,1,1)(1,1,0)[7]             : OOB=20454554011.088, Time=0.98 sec
 ARIMA(1,1,1)(1,1,1)[7]             : OOB=1530781725.977, Time=1.72 sec
 ARIMA(1,1,1)(1,1,2)[7]             : OOB=1577916328.480, Time=2.44 sec
 ARIMA(1,1,1)(1,1,3)[7]             : OOB=1749840888.162, Time=4.10 sec
 ARIMA(1,1,1)(1,1,4)[7]             : OOB=1527874514.510, Time=7.92 sec
 ARIMA(1,1,1)(1,1,5)[7]             : OOB=1811574470.492, Time=7.82 sec
 ARIMA(1,1,1)(2,1,0)[7]             : OOB=21760592751.809, Time=2.08 sec
 ARIMA(1,1,1)(2,1,1)[7]             : OOB=1521981850.379, Time=2.01 sec
 ARIMA(1,1,1)(2,1,2)[7]             : OOB=1539902117.897, Time=2.36 sec
 ARIMA(1,1,1)(2,1,3)[7]             : OOB=1754340975.061, Time=4.08 sec
 ARIMA(1,1,1)(2,1,4)[7]             : OOB=1730609212.044, Time=7.79 sec
 ARIMA(1,1,1)(3,1,0)[7]             : OOB=18434793119.136, Time=4.45 sec
 ARIMA(1,1,1)(3,1,1)[7]             : OOB=1803432287.608, Time=2.87 sec
 ARIMA(1,1,1)(3,1,2)[7]             : OOB=1598885035.005, Time=4.00 sec
 ARIMA(1,1,1)(3,1,3)[7]             : OOB=1514687629.393, Time=3.50 sec
 ARIMA(1,1,1)(4,1,0)[7]             : OOB=14625717366.132, Time=4.15 sec
 ARIMA(1,1,1)(4,1,1)[7]             : OOB=1829108489.730, Time=5.04 sec
 ARIMA(1,1,1)(4,1,2)[7]             : OOB=1695605536.569, Time=5.28 sec
 ARIMA(1,1,1)(5,1,0)[7]             : OOB=12888124451.644, Time=4.39 sec
 ARIMA(1,1,1)(5,1,1)[7]             : OOB=1647636381.313, Time=11.69 sec
 ARIMA(1,1,2)(0,1,0)[7]             : OOB=22104710904.533, Time=4.48 sec
 ARIMA(1,1,2)(0,1,1)[7]             : OOB=10463489094.840, Time=3.38 sec
 ARIMA(1,1,2)(0,1,2)[7]             : OOB=2089827533.255, Time=4.20 sec
 ARIMA(1,1,2)(0,1,3)[7]             : OOB=1656894155.110, Time=3.63 sec
 ARIMA(1,1,2)(0,1,4)[7]             : OOB=1521916663.244, Time=6.75 sec
 ARIMA(1,1,2)(0,1,5)[7]             : OOB=1512350799.101, Time=9.97 sec
 ARIMA(1,1,2)(1,1,0)[7]             : OOB=20426953693.734, Time=2.06 sec
 ARIMA(1,1,2)(1,1,1)[7]             : OOB=1554380852.082, Time=2.15 sec
 ARIMA(1,1,2)(1,1,2)[7]             : OOB=1530491661.447, Time=3.99 sec
 ARIMA(1,1,2)(1,1,3)[7]             : OOB=1798612871.123, Time=5.85 sec
 ARIMA(1,1,2)(1,1,4)[7]             : OOB=1550535398.990, Time=10.38 sec
 ARIMA(1,1,2)(2,1,0)[7]             : OOB=21548879048.894, Time=2.41 sec
 ARIMA(1,1,2)(2,1,1)[7]             : OOB=1552045850.112, Time=1.51 sec
 ARIMA(1,1,2)(2,1,2)[7]             : OOB=inf, Time=4.33 sec
 ARIMA(1,1,2)(2,1,3)[7]             : OOB=1787731895.341, Time=5.93 sec
 ARIMA(1,1,2)(3,1,0)[7]             : OOB=18517477354.202, Time=4.25 sec
 ARIMA(1,1,2)(3,1,1)[7]             : OOB=1966560495.957, Time=3.86 sec
 ARIMA(1,1,2)(3,1,2)[7]             : OOB=1627897776.327, Time=3.87 sec
 ARIMA(1,1,2)(4,1,0)[7]             : OOB=14505674873.555, Time=10.07 sec
 ARIMA(1,1,2)(4,1,1)[7]             : OOB=1825438241.834, Time=24.31 sec
 ARIMA(1,1,2)(5,1,0)[7]             : OOB=12733134110.390, Time=14.09 sec
 ARIMA(1,1,3)(0,1,0)[7]             : OOB=23949389540.561, Time=6.93 sec
 ARIMA(1,1,3)(0,1,1)[7]             : OOB=12065965278.142, Time=8.96 sec
 ARIMA(1,1,3)(0,1,2)[7]             : OOB=1873455195.144, Time=5.62 sec
 ARIMA(1,1,3)(0,1,3)[7]             : OOB=1557489266.831, Time=12.10 sec
 ARIMA(1,1,3)(0,1,4)[7]             : OOB=1396319590.621, Time=11.48 sec
 ARIMA(1,1,3)(1,1,0)[7]             : OOB=21805928015.556, Time=3.95 sec
 ARIMA(1,1,3)(1,1,1)[7]             : OOB=1406081034.900, Time=3.39 sec
 ARIMA(1,1,3)(1,1,2)[7]             : OOB=1403218111.876, Time=3.34 sec
 ARIMA(1,1,3)(1,1,3)[7]             : OOB=1640096383.378, Time=8.86 sec
 ARIMA(1,1,3)(2,1,0)[7]             : OOB=22984502190.722, Time=2.54 sec
 ARIMA(1,1,3)(2,1,1)[7]             : OOB=1403487163.466, Time=4.20 sec
 ARIMA(1,1,3)(2,1,2)[7]             : OOB=1411184178.142, Time=3.95 sec
 ARIMA(1,1,3)(3,1,0)[7]             : OOB=19712600287.743, Time=5.62 sec
 ARIMA(1,1,3)(3,1,1)[7]             : OOB=1851996311.604, Time=6.44 sec
 ARIMA(1,1,3)(4,1,0)[7]             : OOB=15482875811.002, Time=8.47 sec
 ARIMA(1,1,4)(0,1,0)[7]             : OOB=23896582781.662, Time=6.29 sec
 ARIMA(1,1,4)(0,1,1)[7]             : OOB=12158150993.745, Time=4.33 sec
 ARIMA(1,1,4)(0,1,2)[7]             : OOB=1953945390.644, Time=4.21 sec
 ARIMA(1,1,4)(0,1,3)[7]             : OOB=1594847898.477, Time=13.29 sec
 ARIMA(1,1,4)(1,1,0)[7]             : OOB=21679215828.793, Time=6.73 sec
 ARIMA(1,1,4)(1,1,1)[7]             : OOB=1444898978.960, Time=4.47 sec
 ARIMA(1,1,4)(1,1,2)[7]             : OOB=1439043577.601, Time=5.88 sec
 ARIMA(1,1,4)(2,1,0)[7]             : OOB=22948988672.017, Time=11.74 sec
 ARIMA(1,1,4)(2,1,1)[7]             : OOB=1440321844.018, Time=4.35 sec
 ARIMA(1,1,4)(3,1,0)[7]             : OOB=19621353273.085, Time=5.15 sec
 ARIMA(1,1,5)(0,1,0)[7]             : OOB=23628852694.675, Time=3.52 sec
 ARIMA(1,1,5)(0,1,1)[7]             : OOB=12182629340.486, Time=2.57 sec
 ARIMA(1,1,5)(0,1,2)[7]             : OOB=1946687807.083, Time=3.72 sec
 ARIMA(1,1,5)(1,1,0)[7]             : OOB=21911863863.879, Time=1.24 sec
 ARIMA(1,1,5)(1,1,1)[7]             : OOB=1412364545.705, Time=1.44 sec
 ARIMA(1,1,5)(2,1,0)[7]             : OOB=22727910050.466, Time=2.08 sec
 ARIMA(2,1,0)(0,1,0)[7]             : OOB=18791795716.170, Time=2.27 sec
 ARIMA(2,1,0)(0,1,1)[7]             : OOB=8860328836.258, Time=1.55 sec
 ARIMA(2,1,0)(0,1,2)[7]             : OOB=2159057680.997, Time=1.82 sec
 ARIMA(2,1,0)(0,1,3)[7]             : OOB=1685914056.086, Time=5.74 sec
 ARIMA(2,1,0)(0,1,4)[7]             : OOB=1550940314.641, Time=534.72 sec
 ARIMA(2,1,0)(0,1,5)[7]             : OOB=1563379851.913, Time=5.47 sec
 ARIMA(2,1,0)(1,1,0)[7]             : OOB=17778317074.448, Time=1.17 sec
 ARIMA(2,1,0)(1,1,1)[7]             : OOB=1586704476.928, Time=1.51 sec
 ARIMA(2,1,0)(1,1,2)[7]             : OOB=1574424425.355, Time=2.47 sec
 ARIMA(2,1,0)(1,1,3)[7]             : OOB=1703248934.919, Time=3.76 sec
 ARIMA(2,1,0)(1,1,4)[7]             : OOB=1580414166.257, Time=5.62 sec
 ARIMA(2,1,0)(1,1,5)[7]             : OOB=1550958518.180, Time=5.74 sec
 ARIMA(2,1,0)(2,1,0)[7]             : OOB=18560142074.003, Time=1.26 sec
 ARIMA(2,1,0)(2,1,1)[7]             : OOB=1561681062.838, Time=1.40 sec
 ARIMA(2,1,0)(2,1,2)[7]             : OOB=1595405838.997, Time=1.83 sec
 ARIMA(2,1,0)(2,1,3)[7]             : OOB=1659574232.514, Time=3.06 sec
 ARIMA(2,1,0)(2,1,4)[7]             : OOB=1659161838.981, Time=7.42 sec
 ARIMA(2,1,0)(3,1,0)[7]             : OOB=15888191358.835, Time=1.50 sec
 ARIMA(2,1,0)(3,1,1)[7]             : OOB=1701030998.732, Time=2.02 sec
 ARIMA(2,1,0)(3,1,2)[7]             : OOB=1579794452.589, Time=2.67 sec
 ARIMA(2,1,0)(3,1,3)[7]             : OOB=inf, Time=6.56 sec
 ARIMA(2,1,0)(4,1,0)[7]             : OOB=12450409247.084, Time=3.62 sec
 ARIMA(2,1,0)(4,1,1)[7]             : OOB=1765442238.884, Time=5.19 sec
 ARIMA(2,1,0)(4,1,2)[7]             : OOB=1607395894.454, Time=5.53 sec
 ARIMA(2,1,0)(5,1,0)[7]             : OOB=11099268908.717, Time=5.02 sec
 ARIMA(2,1,0)(5,1,1)[7]             : OOB=1706667488.562, Time=9.03 sec
 ARIMA(2,1,1)(0,1,0)[7]             : OOB=17879344870.231, Time=2.50 sec
 ARIMA(2,1,1)(0,1,1)[7]             : OOB=10595392017.367, Time=1.27 sec
 ARIMA(2,1,1)(0,1,2)[7]             : OOB=2167391937.800, Time=1.66 sec
 ARIMA(2,1,1)(0,1,3)[7]             : OOB=1608277330.721, Time=3.13 sec
 ARIMA(2,1,1)(0,1,4)[7]             : OOB=1454998590.261, Time=4.43 sec
 ARIMA(2,1,1)(0,1,5)[7]             : OOB=1487257054.845, Time=5.68 sec
 ARIMA(2,1,1)(1,1,0)[7]             : OOB=20975533129.958, Time=0.96 sec
 ARIMA(2,1,1)(1,1,1)[7]             : OOB=1492011181.718, Time=2.17 sec
 ARIMA(2,1,1)(1,1,2)[7]             : OOB=1524452938.447, Time=2.58 sec
 ARIMA(2,1,1)(1,1,3)[7]             : OOB=1691654479.414, Time=3.62 sec
 ARIMA(2,1,1)(1,1,4)[7]             : OOB=1497805855.067, Time=6.67 sec
 ARIMA(2,1,1)(2,1,0)[7]             : OOB=21875391617.169, Time=1.77 sec
 ARIMA(2,1,1)(2,1,1)[7]             : OOB=1482929199.814, Time=2.72 sec
 ARIMA(2,1,1)(2,1,2)[7]             : OOB=1498859633.364, Time=1.81 sec
 ARIMA(2,1,1)(2,1,3)[7]             : OOB=1689479481.845, Time=2.88 sec
 ARIMA(2,1,1)(3,1,0)[7]             : OOB=18566651817.578, Time=2.65 sec
 ARIMA(2,1,1)(3,1,1)[7]             : OOB=1748783212.912, Time=3.22 sec
 ARIMA(2,1,1)(3,1,2)[7]             : OOB=1549130457.455, Time=3.01 sec
 ARIMA(2,1,1)(4,1,0)[7]             : OOB=14651118637.776, Time=4.65 sec
 ARIMA(2,1,1)(4,1,1)[7]             : OOB=1741767380.936, Time=7.63 sec
 ARIMA(2,1,1)(5,1,0)[7]             : OOB=12815378445.872, Time=7.36 sec
 ARIMA(2,1,2)(0,1,0)[7]             : OOB=23828204746.576, Time=4.44 sec
 ARIMA(2,1,2)(0,1,1)[7]             : OOB=11988373250.380, Time=1.90 sec
 ARIMA(2,1,2)(0,1,2)[7]             : OOB=1901189071.364, Time=2.47 sec
 ARIMA(2,1,2)(0,1,3)[7]             : OOB=1574921170.506, Time=4.12 sec
 ARIMA(2,1,2)(0,1,4)[7]             : OOB=1401832596.626, Time=6.97 sec
 ARIMA(2,1,2)(1,1,0)[7]             : OOB=21699331492.950, Time=1.26 sec
 ARIMA(2,1,2)(1,1,1)[7]             : OOB=1414463587.059, Time=2.21 sec
 ARIMA(2,1,2)(1,1,2)[7]             : OOB=1412849079.108, Time=3.15 sec
 ARIMA(2,1,2)(1,1,3)[7]             : OOB=1647871366.360, Time=5.24 sec
 ARIMA(2,1,2)(2,1,0)[7]             : OOB=22797962018.948, Time=3.15 sec
 ARIMA(2,1,2)(2,1,1)[7]             : OOB=1413379437.694, Time=2.22 sec
 ARIMA(2,1,2)(2,1,2)[7]             : OOB=1419666356.967, Time=3.60 sec
 ARIMA(2,1,2)(3,1,0)[7]             : OOB=19551102216.390, Time=3.09 sec
 ARIMA(2,1,2)(3,1,1)[7]             : OOB=1848482441.724, Time=4.14 sec
 ARIMA(2,1,2)(4,1,0)[7]             : OOB=15342119782.190, Time=4.98 sec
 ARIMA(2,1,3)(0,1,0)[7]             : OOB=inf, Time=2.82 sec
 ARIMA(2,1,3)(0,1,1)[7]             : OOB=inf, Time=4.79 sec
 ARIMA(2,1,3)(0,1,2)[7]             : OOB=1959113883.286, Time=4.80 sec
 ARIMA(2,1,3)(0,1,3)[7]             : OOB=1562593283.867, Time=7.62 sec
 ARIMA(2,1,3)(1,1,0)[7]             : OOB=inf, Time=2.18 sec
 ARIMA(2,1,3)(1,1,1)[7]             : OOB=1449321294.945, Time=2.34 sec
 ARIMA(2,1,3)(1,1,2)[7]             : OOB=1432554138.140, Time=3.86 sec
 ARIMA(2,1,3)(2,1,0)[7]             : OOB=22883945139.311, Time=3.30 sec
 ARIMA(2,1,3)(2,1,1)[7]             : OOB=1435397118.504, Time=3.39 sec
 ARIMA(2,1,3)(3,1,0)[7]             : OOB=19539929084.364, Time=4.76 sec
 ARIMA(2,1,4)(0,1,0)[7]             : OOB=inf, Time=3.11 sec
 ARIMA(2,1,4)(0,1,1)[7]             : OOB=9807105798.702, Time=3.40 sec
 ARIMA(2,1,4)(0,1,2)[7]             : OOB=4600525000.385, Time=3.66 sec
 ARIMA(2,1,4)(1,1,0)[7]             : OOB=inf, Time=2.11 sec
 ARIMA(2,1,4)(1,1,1)[7]             : OOB=1442955141.495, Time=2.29 sec
 ARIMA(2,1,4)(2,1,0)[7]             : OOB=22884023943.580, Time=3.24 sec
 ARIMA(2,1,5)(0,1,0)[7]             : OOB=inf, Time=3.64 sec
 ARIMA(2,1,5)(0,1,1)[7]             : OOB=9212297843.579, Time=5.38 sec
 ARIMA(2,1,5)(1,1,0)[7]             : OOB=inf, Time=2.44 sec
 ARIMA(3,1,0)(0,1,0)[7]             : OOB=22749576743.906, Time=0.63 sec
 ARIMA(3,1,0)(0,1,1)[7]             : OOB=10687820354.494, Time=0.68 sec
 ARIMA(3,1,0)(0,1,2)[7]             : OOB=2190184039.296, Time=0.97 sec
 ARIMA(3,1,0)(0,1,3)[7]             : OOB=1673148532.971, Time=1.40 sec
 ARIMA(3,1,0)(0,1,4)[7]             : OOB=1540014008.304, Time=2.84 sec
 ARIMA(3,1,0)(0,1,5)[7]             : OOB=1593572377.097, Time=3.44 sec
 ARIMA(3,1,0)(1,1,0)[7]             : OOB=20536257042.225, Time=0.61 sec
 ARIMA(3,1,0)(1,1,1)[7]             : OOB=1558981610.440, Time=1.27 sec
 ARIMA(3,1,0)(1,1,2)[7]             : OOB=1580801483.932, Time=1.68 sec
 ARIMA(3,1,0)(1,1,3)[7]             : OOB=1812965359.896, Time=2.60 sec
 ARIMA(3,1,0)(1,1,4)[7]             : OOB=1547208216.619, Time=4.86 sec
 ARIMA(3,1,0)(2,1,0)[7]             : OOB=21901824304.257, Time=1.04 sec
 ARIMA(3,1,0)(2,1,1)[7]             : OOB=1563815107.562, Time=1.60 sec
 ARIMA(3,1,0)(2,1,2)[7]             : OOB=1576451238.609, Time=1.81 sec
 ARIMA(3,1,0)(2,1,3)[7]             : OOB=1789533168.403, Time=2.74 sec
 ARIMA(3,1,0)(3,1,0)[7]             : OOB=18654238418.098, Time=1.73 sec
 ARIMA(3,1,0)(3,1,1)[7]             : OOB=1892579738.170, Time=2.76 sec
 ARIMA(3,1,0)(3,1,2)[7]             : OOB=1665301804.917, Time=3.40 sec
 ARIMA(3,1,0)(4,1,0)[7]             : OOB=14722138221.436, Time=2.96 sec
 ARIMA(3,1,0)(4,1,1)[7]             : OOB=1828803674.307, Time=4.96 sec
 ARIMA(3,1,0)(5,1,0)[7]             : OOB=12913819361.488, Time=5.89 sec
 ARIMA(3,1,1)(0,1,0)[7]             : OOB=21803364831.832, Time=1.88 sec
 ARIMA(3,1,1)(0,1,1)[7]             : OOB=10844421257.519, Time=1.35 sec
 ARIMA(3,1,1)(0,1,2)[7]             : OOB=2057489507.492, Time=2.15 sec
 ARIMA(3,1,1)(0,1,3)[7]             : OOB=1564315925.786, Time=2.07 sec
 ARIMA(3,1,1)(0,1,4)[7]             : OOB=1417324179.754, Time=3.61 sec
 ARIMA(3,1,1)(1,1,0)[7]             : OOB=21070756689.438, Time=1.67 sec
 ARIMA(3,1,1)(1,1,1)[7]             : OOB=1451839428.964, Time=1.38 sec
 ARIMA(3,1,1)(1,1,2)[7]             : OOB=1485162722.575, Time=2.39 sec
 ARIMA(3,1,1)(1,1,3)[7]             : OOB=1671814370.932, Time=3.51 sec
 ARIMA(3,1,1)(2,1,0)[7]             : OOB=22512762262.651, Time=2.03 sec
 ARIMA(3,1,1)(2,1,1)[7]             : OOB=1453930992.187, Time=3.29 sec
 ARIMA(3,1,1)(2,1,2)[7]             : OOB=1461953770.284, Time=2.96 sec
 ARIMA(3,1,1)(3,1,0)[7]             : OOB=19113937467.820, Time=2.82 sec
 ARIMA(3,1,1)(3,1,1)[7]             : OOB=1762748890.545, Time=5.95 sec
 ARIMA(3,1,1)(4,1,0)[7]             : OOB=13984362794.713, Time=8.25 sec
 ARIMA(3,1,2)(0,1,0)[7]             : OOB=21852258215.742, Time=2.30 sec
 ARIMA(3,1,2)(0,1,1)[7]             : OOB=12080957361.953, Time=1.86 sec
 ARIMA(3,1,2)(0,1,2)[7]             : OOB=1929766876.471, Time=2.39 sec
 ARIMA(3,1,2)(0,1,3)[7]             : OOB=1586930739.477, Time=5.27 sec
 ARIMA(3,1,2)(1,1,0)[7]             : OOB=21619669531.735, Time=3.75 sec
 ARIMA(3,1,2)(1,1,1)[7]             : OOB=1432689260.978, Time=2.28 sec
 ARIMA(3,1,2)(1,1,2)[7]             : OOB=1428753724.670, Time=3.29 sec
 ARIMA(3,1,2)(2,1,0)[7]             : OOB=22911893299.275, Time=1.95 sec
 ARIMA(3,1,2)(2,1,1)[7]             : OOB=1428813233.550, Time=2.62 sec
 ARIMA(3,1,2)(3,1,0)[7]             : OOB=19594139157.825, Time=4.19 sec
 ARIMA(3,1,3)(0,1,0)[7]             : OOB=inf, Time=4.03 sec
 ARIMA(3,1,3)(0,1,1)[7]             : OOB=7366329626.989, Time=3.00 sec
 ARIMA(3,1,3)(0,1,2)[7]             : OOB=4476276708.831, Time=3.70 sec
 ARIMA(3,1,3)(1,1,0)[7]             : OOB=inf, Time=2.54 sec
 ARIMA(3,1,3)(1,1,1)[7]             : OOB=1502577125.821, Time=2.92 sec
 ARIMA(3,1,3)(2,1,0)[7]             : OOB=23126321269.115, Time=2.10 sec
 ARIMA(3,1,4)(0,1,0)[7]             : OOB=inf, Time=5.51 sec
 ARIMA(3,1,4)(0,1,1)[7]             : OOB=inf, Time=4.73 sec
 ARIMA(3,1,4)(1,1,0)[7]             : OOB=inf, Time=3.41 sec
 ARIMA(3,1,5)(0,1,0)[7]             : OOB=inf, Time=8.36 sec
 ARIMA(4,1,0)(0,1,0)[7]             : OOB=22691907781.876, Time=1.16 sec
 ARIMA(4,1,0)(0,1,1)[7]             : OOB=10714254719.844, Time=1.06 sec
 ARIMA(4,1,0)(0,1,2)[7]             : OOB=2182901069.474, Time=1.87 sec
 ARIMA(4,1,0)(0,1,3)[7]             : OOB=1655708186.995, Time=2.22 sec
 ARIMA(4,1,0)(0,1,4)[7]             : OOB=1517885601.941, Time=5.81 sec
 ARIMA(4,1,0)(1,1,0)[7]             : OOB=20587495241.247, Time=2.35 sec
 ARIMA(4,1,0)(1,1,1)[7]             : OOB=1540777316.499, Time=1.27 sec
 ARIMA(4,1,0)(1,1,2)[7]             : OOB=1564534014.391, Time=3.88 sec
 ARIMA(4,1,0)(1,1,3)[7]             : OOB=1788007807.177, Time=2.83 sec
 ARIMA(4,1,0)(2,1,0)[7]             : OOB=21887588803.430, Time=1.35 sec
 ARIMA(4,1,0)(2,1,1)[7]             : OOB=1544768082.105, Time=5.40 sec
 ARIMA(4,1,0)(2,1,2)[7]             : OOB=1559817115.053, Time=2.56 sec
 ARIMA(4,1,0)(3,1,0)[7]             : OOB=18579893035.437, Time=2.75 sec
 ARIMA(4,1,0)(3,1,1)[7]             : OOB=1871168180.957, Time=4.78 sec
 ARIMA(4,1,0)(4,1,0)[7]             : OOB=14620551157.417, Time=4.61 sec
 ARIMA(4,1,1)(0,1,0)[7]             : OOB=inf, Time=3.77 sec
 ARIMA(4,1,1)(0,1,1)[7]             : OOB=10931175909.788, Time=0.98 sec
 ARIMA(4,1,1)(0,1,2)[7]             : OOB=2093600700.362, Time=1.39 sec
 ARIMA(4,1,1)(0,1,3)[7]             : OOB=1585875234.616, Time=2.48 sec
 ARIMA(4,1,1)(1,1,0)[7]             : OOB=inf, Time=2.22 sec
 ARIMA(4,1,1)(1,1,1)[7]             : OOB=1469258797.290, Time=1.58 sec
 ARIMA(4,1,1)(1,1,2)[7]             : OOB=1498845157.507, Time=3.88 sec
 ARIMA(4,1,1)(2,1,0)[7]             : OOB=22520157668.336, Time=2.52 sec
 ARIMA(4,1,1)(2,1,1)[7]             : OOB=1470155439.120, Time=4.76 sec
 ARIMA(4,1,1)(3,1,0)[7]             : OOB=19119016222.992, Time=6.50 sec
 ARIMA(4,1,2)(0,1,0)[7]             : OOB=inf, Time=3.64 sec
 ARIMA(4,1,2)(0,1,1)[7]             : OOB=12084961817.690, Time=2.73 sec
 ARIMA(4,1,2)(0,1,2)[7]             : OOB=1929258769.489, Time=2.92 sec
 ARIMA(4,1,2)(1,1,0)[7]             : OOB=inf, Time=4.33 sec
 ARIMA(4,1,2)(1,1,1)[7]             : OOB=1430598503.952, Time=2.83 sec
 ARIMA(4,1,2)(2,1,0)[7]             : OOB=22699642678.089, Time=5.49 sec
 ARIMA(4,1,3)(0,1,0)[7]             : OOB=inf, Time=3.60 sec
 ARIMA(4,1,3)(0,1,1)[7]             : OOB=8178219627.625, Time=3.89 sec
 ARIMA(4,1,3)(1,1,0)[7]             : OOB=inf, Time=2.46 sec
 ARIMA(4,1,4)(0,1,0)[7]             : OOB=inf, Time=4.78 sec
 ARIMA(5,1,0)(0,1,0)[7]             : OOB=23493647127.455, Time=2.23 sec
 ARIMA(5,1,0)(0,1,1)[7]             : OOB=10899210803.059, Time=1.22 sec
 ARIMA(5,1,0)(0,1,2)[7]             : OOB=1845977466.296, Time=2.26 sec
 ARIMA(5,1,0)(0,1,3)[7]             : OOB=1502782973.624, Time=2.29 sec
 ARIMA(5,1,0)(1,1,0)[7]             : OOB=21448462505.682, Time=1.36 sec
 ARIMA(5,1,0)(1,1,1)[7]             : OOB=1374269331.531, Time=1.78 sec
 ARIMA(5,1,0)(1,1,2)[7]             : OOB=1383092696.425, Time=1.94 sec
 ARIMA(5,1,0)(2,1,0)[7]             : OOB=22994522549.243, Time=2.60 sec
 ARIMA(5,1,0)(2,1,1)[7]             : OOB=1368304885.419, Time=2.09 sec
 ARIMA(5,1,0)(3,1,0)[7]             : OOB=19636375627.412, Time=3.24 sec
 ARIMA(5,1,1)(0,1,0)[7]             : OOB=23006201133.475, Time=4.65 sec
 ARIMA(5,1,1)(0,1,1)[7]             : OOB=10773225250.536, Time=1.96 sec
 ARIMA(5,1,1)(0,1,2)[7]             : OOB=1862494384.327, Time=2.70 sec
 ARIMA(5,1,1)(1,1,0)[7]             : OOB=21182236065.380, Time=1.83 sec
 ARIMA(5,1,1)(1,1,1)[7]             : OOB=1365766667.637, Time=1.52 sec
 ARIMA(5,1,1)(2,1,0)[7]             : OOB=22765588455.097, Time=2.55 sec
 ARIMA(5,1,2)(0,1,0)[7]             : OOB=inf, Time=3.54 sec
 ARIMA(5,1,2)(0,1,1)[7]             : OOB=4991305677.625, Time=2.17 sec
 ARIMA(5,1,2)(1,1,0)[7]             : OOB=inf, Time=3.07 sec
 ARIMA(5,1,3)(0,1,0)[7]             : OOB=inf, Time=2.64 sec

Best model:  ARIMA(5,1,1)(1,1,1)[7]          
Total fit time: 2397.948 seconds

ARIMA(maxiter=200, order=(5, 1, 1), out_of_sample_size=30, scoring_args={},
      seasonal_order=(1, 1, 1, 7), suppress_warnings=True,
      with_intercept=False)


#Calculating RMSE after 1 month daily rolling forecast on the best selected model above by auto_arima

train_and_pred(((5,1,1),(1,1,1,7)), aic_bic=False).rmse()

23470.31509631157


#Top performing models from auto_arima and some manually selected neighbouring models balancing AR and MA

sarimax_top_models = [((5,0,1),(1,0,0,7)),
                      ((1,1,0),(1,0,0,7)),
                      ((5,0,0),(1,0,0,7)),
                      ((1,0,0),(1,0,0,7)),
                      ((0,1,0),(2,0,1,7)),
                      ((1,1,0),(0,1,0,7)),
                      ((1,1,0),(1,1,0,7)),
                      ((1,0,0),(4,0,0,7)),
                      ((0,1,0),(1,0,5,7)),
                      ((5,0,1),(1,1,0,7)),
                      ((3,1,3),(1,0,0,7)),
                      ((5,0,1),(1,0,1,7)),
                      ((5,1,4),(0,1,0,7)),
                      ((0,1,0),(2,0,1,7)),
                      ((2,1,0),(3,1,0,7)),
                      ((2,0,2),(1,0,0,7))]


%%time
#Training top models on (Train+Validation) set and predicting with rolling forecast on 2nd Validation set
sarimax_top_val2 = {'normal':[],
                    'boxcox':[],
                    'lr':[]}

#Top selected models from auto_arima
for x in tqdm(sarimax_top_models):
    a = train_and_pred(x, aic_bic=False, split_type='val2')
    sarimax_top_val2['normal'].append(a)
    b = train_and_pred(x, aic_bic=False, split_type='val2', series_type='boxcox')
    sarimax_top_val2['boxcox'].append(b)
    c = train_and_pred(x, aic_bic=False, split_type='val2', series_type='lr')
    sarimax_top_val2['lr'].append(c)

#Top models from grid search
for x in tqdm(top_models_grid[top_models_grid[2] == 'normal'][0].values):
    a = train_and_pred(x, aic_bic=False, split_type='val2')
    sarimax_top_val2['normal'].append(a)
for x in tqdm(top_models_grid[top_models_grid[2] == 'boxcox'][0].values):
    a = train_and_pred(x, aic_bic=False, split_type='val2', series_type='boxcox')
    sarimax_top_val2['boxcox'].append(a)
for x in tqdm(top_models_grid[top_models_grid[2] == 'lr'][0].values):
    a = train_and_pred(x, aic_bic=False, split_type='val2', series_type='lr')
    sarimax_top_val2['lr'].append(a)

100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [21:43<00:00, 81.44s/it]
100%|███████████████████████████████████████████████████████████████████████████████████| 9/9 [56:56<00:00, 379.65s/it]
100%|███████████████████████████████████████████████████████████████████████████████| 27/27 [1:38:14<00:00, 218.31s/it]
100%|███████████████████████████████████████████████████████████████████████████████| 14/14 [1:15:39<00:00, 324.26s/it]

Wall time: 4h 12min 34s


%%time
#Calculating rolling forecast RMSE scores of top models on 1st Validation set, for comparison. (20~ MINS TO RUN)

sarimax_top_val = {'normal':[],
                    'boxcox':[],
                    'lr':[]}

for x in tqdm(sarimax_top_models):
    a = train_and_pred(x, aic_bic=False, split_type='val')
    sarimax_top_val['normal'].append(a)
    b = train_and_pred(x, aic_bic=False, split_type='val', series_type='boxcox')
    sarimax_top_val['boxcox'].append(b)
    c = train_and_pred(x, aic_bic=False, split_type='val', series_type='lr')
    sarimax_top_val['lr'].append(c)

100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [20:03<00:00, 75.24s/it]

Wall time: 20min 42s


#Removing None from results, results with extreme outliers, results with missing predicted values

sarimax_top_val2['normal'] = [x for x in sarimax_top_val2['normal'] if x!=None]
sarimax_top_val2['boxcox'] = [x for x in sarimax_top_val2['boxcox'] if x!=None]
sarimax_top_val2['lr'] = [x for x in sarimax_top_val2['lr'] if x!=None]


#Rolling forecast RMSE scores of top models on normal, boxcox and LR series

sarima_top_list = [(x.model_attribs, x.rmse(),'normal') for x in sarimax_top_val2['normal'] if x!=None]
sarima_top_list += [(x.model_attribs, x.rmse(),'boxcox') for x in sarimax_top_val2['boxcox'] if x!=None]
sarima_top_list += [(x.model_attribs, x.rmse(),'lr') for x in sarimax_top_val2['lr'] if x!= None]


#Results sorted from best to worst by RMSE scores

pd.DataFrame(sarima_top_list).sort_values(1)


#Making dataframes of top models on each series type along with model scores and appending them to a list

a = []
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'normal','val2') for x in sarimax_top_val2['normal'] if x!=None]))
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'boxcox','val2') for x in sarimax_top_val2['boxcox'] if x!=None]))
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'lr','val2') for x in sarimax_top_val2['lr'] if x!=None]))
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'normal','val') for x in sarimax_top_val['normal'] if x!=None]))
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'boxcox','val') for x in sarimax_top_val['boxcox'] if x!=None]))
a.append(pd.DataFrame([(x.model_attribs, x.rmse(), 'lr','val') for x in sarimax_top_val['lr'] if x!=None]))
top_models_grid[3] = 'val'
a.append(top_models_grid)


top_trained_val2 = train_and_pred(((5, 0, 1), (1, 1, 0, 7)), aic_bic=False, split_type='val2', series_type='boxcox')
top_trained_val2.fig()


top_trained_val2.resid_fig()


top_trained_val2.get_all()

For model ((5, 0, 1), (1, 1, 0, 7)):

Mean Squared Error (MSE):  821830219.4416555 
Root Mean Squared Error (RMSE):  28667.58133225849 
Mean Absolute Error (MAE):  21563.684736997533 
Mean Absolute Percentage Error (MAPE):  4.2061788801090145 
Mean Absolute Scaled Error (MASE):  236.7215675662391


#Formatting scores and calculating average to select best generalising model

scores_df = pd.concat(a, ignore_index=True)
scores_df['val1'] = scores_df[scores_df[3] == 'val'][1]
scores_df['val2'] = scores_df[scores_df[3] == 'val2'][1]
scores_df.drop([1,3], axis=1, inplace=True)
scores_df = scores_df.groupby([0,2]).last().reset_index()
scores_df.columns = ['Model','Series Type','Validation 1','Validation 2']
scores_df['Avg'] = (scores_df['Validation 1'] + scores_df['Validation 2'])/2
scores_df['Diff'] = np.abs(scores_df['Validation 1'] - scores_df['Validation 2'])
scores_df = scores_df.sort_values('Avg')
scores_df = scores_df.round(2)


fig = go.Figure(data=[go.Table(
    header=dict(values=list(scores_df.columns),
                fill_color='paleturquoise',
                align='left'),
    cells=dict(values=[scores_df.Model, scores_df['Series Type'],scores_df['Validation 1'],scores_df['Validation 2'],
                      scores_df['Avg'], scores_df['Diff']],
               fill_color='lavender',
               align='left'))
])
fig.update_layout(title_text='RMSE Scores SARIMA',margin=dict(b=0), height=700)
fig.show(config=config)


#Training best generalising model till now to predict on the test set

test_final = train_and_pred(((1, 1, 0), (1, 1, 0, 7)), aic_bic=False, series_type='boxcox', split_type='test')


test_final.fig()


test_final.resid_fig()


test_final.get_all()

For model ((1, 1, 0), (1, 1, 0, 7)):

Mean Squared Error (MSE):  973344399.0106475 
Root Mean Squared Error (RMSE):  31198.467895245234 
Mean Absolute Error (MAE):  23504.886238949755 
Mean Absolute Percentage Error (MAPE):  3.8878369817747185 
Mean Absolute Scaled Error (MASE):  174.79105800746692


#Creating an ensemble of the best models in a list

scores_df.reset_index(drop=True, inplace=True)
test_ensemble = [(scores_df.head(50).groupby(['Series Type','Model']).last()).loc['normal'].index.values,
                 (scores_df.head(50).groupby(['Series Type','Model']).last()).loc['boxcox'].index.values,
                 (scores_df.head(50).groupby(['Series Type','Model']).last()).loc['lr'].index.values]


%%time
#Training the models in the ensemble list to predict on Test set

sarimax_top_test = {'normal':[],
                    'boxcox':[],
                    'lr':[]}
for x in test_ensemble[0]:
    a = train_and_pred(x, aic_bic=False, split_type='test')
    sarimax_top_test['normal'].append(a)
for x in test_ensemble[1]:
    b = train_and_pred(x, aic_bic=False, split_type='test', series_type='boxcox')
    sarimax_top_test['boxcox'].append(b)
for x in test_ensemble[2]:
    c = train_and_pred(x, aic_bic=False, split_type='test', series_type='lr')
    sarimax_top_test['lr'].append(c)

Wall time: 1h 39min 9s


#Models with their RMSE scores on Test set on normal series

[(x.model_attribs, x.rmse()) for x in sarimax_top_test['normal'] if x!=None]

[(((1, 1, 0), (0, 1, 0, 7)), 29857.16126774989),
 (((1, 1, 0), (1, 1, 0, 7)), 33154.58170657405),
 (((2, 1, 0), (3, 1, 0, 7)), 38153.19755293955),
 (((3, 1, 3), (1, 0, 0, 7)), 37612.440730236514),
 (((5, 0, 1), (1, 1, 0, 7)), 30860.777083694804),
 (((5, 1, 3), (3, 0, 0, 7)), 41955.9465115121),
 (((5, 1, 4), (0, 1, 0, 7)), 41008.45526817476)]


#Models with their RMSE scores on Test set on boxcox transformed series

[(x.model_attribs, x.rmse()) for x in sarimax_top_test['boxcox'] if (x!=None and ~np.isnan(x.predicted).any())]

[(((0, 1, 1), (0, 1, 2, 7)), 42598.24434595003),
 (((0, 1, 1), (1, 1, 1, 7)), 44198.24551681934),
 (((1, 0, 1), (0, 1, 2, 7)), 41563.27256085922),
 (((1, 1, 0), (0, 1, 0, 7)), 28136.743096672348),
 (((1, 1, 0), (1, 1, 0, 7)), 31198.467895245234),
 (((1, 1, 3), (0, 1, 2, 7)), 43026.13701682949),
 (((1, 1, 3), (1, 1, 1, 7)), 44643.51639614864),
 (((2, 0, 2), (1, 0, 0, 7)), 37214.832119638115),
 (((2, 0, 2), (1, 1, 1, 7)), 43191.297412920954),
 (((2, 0, 3), (1, 1, 1, 7)), 43462.1386510204),
 (((2, 1, 0), (3, 1, 0, 7)), 36527.769289255855),
 (((2, 1, 3), (1, 1, 1, 7)), 44313.713561726334),
 (((3, 0, 3), (1, 1, 1, 7)), 43914.716774657114),
 (((3, 0, 3), (2, 1, 5, 7)), 46051.83880304545),
 (((3, 1, 3), (1, 0, 0, 7)), 36657.69958515607),
 (((5, 0, 1), (1, 1, 0, 7)), 29146.577899487856),
 (((5, 0, 3), (1, 1, 1, 7)), 44992.062170076606),
 (((5, 1, 3), (0, 1, 2, 7)), 44674.30282273917),
 (((5, 1, 3), (0, 1, 3, 7)), 44908.052547910986),
 (((5, 1, 3), (1, 0, 1, 7)), 42758.864024007344),
 (((5, 1, 3), (1, 0, 2, 7)), 47742.886283542386),
 (((5, 1, 3), (1, 1, 1, 7)), 45067.25807995298),
 (((5, 1, 3), (1, 1, 2, 7)), 44298.21842857938),
 (((5, 1, 3), (2, 0, 1, 7)), 48823.91790832125),
 (((5, 1, 3), (2, 0, 2, 7)), 47065.73935932613),
 (((5, 1, 3), (2, 1, 1, 7)), 45085.726429767004),
 (((5, 1, 3), (3, 0, 1, 7)), 46713.79204760879),
 (((5, 1, 3), (3, 0, 2, 7)), 91506.81261471446),
 (((5, 1, 4), (0, 1, 0, 7)), 42611.26830785137)]


#Models with their RMSE scores on Test set on LR detrended series

[(x.model_attribs, x.rmse()) for x in sarimax_top_test['lr'] if x!=None]

[(((1, 1, 0), (0, 1, 0, 7)), 29857.55026489755),
 (((1, 1, 0), (1, 1, 0, 7)), 33154.34761753777),
 (((2, 0, 2), (1, 0, 0, 7)), 37946.15896420241),
 (((2, 1, 0), (3, 1, 0, 7)), 38153.02488494412),
 (((3, 1, 3), (1, 0, 0, 7)), 37984.09347276074),
 (((5, 0, 1), (1, 0, 0, 7)), 58563.24004942744),
 (((5, 0, 1), (1, 0, 1, 7)), 41195.38050472292),
 (((5, 0, 1), (1, 1, 0, 7)), 30931.89874943104),
 (((5, 1, 3), (1, 0, 1, 7)), 42660.56972307781),
 (((5, 1, 3), (3, 0, 0, 7)), 42011.3938471702),
 (((5, 1, 4), (0, 1, 0, 7)), 41173.72889631861)]


#Taking predictions of all the models in a list

all_pred = []
for key,value in sarimax_top_test.items():
    for x in value:
        if x == None:
            continue
        if np.isnan(x.predicted).any():
            continue
        all_pred.append(x.predicted)


#Saving manually in case the time-taking computation was SKIPPED

all_pred = [[538169.296339039,  472647.37845408206,  576275.302895467,  649669.8245450011,  678715.184368176,  679636.3622903939,  769004.9061356927,  571931.2269493527,  520746.886297117,  588627.3728623524,  669731.1369143045,  700168.0598836961,  704609.925047442,  835811.9285784822,  568144.1576755593,  448143.228845646,  664014.272347302,  649699.6102897462,  724834.065246637,  696429.116936515,  873345.6794873938,  530716.0100915818,  463655.9049312095,  664956.18606667,  666141.6915140856,  723494.1497843121,  700118.4645595823,  873237.5589850979,  520021.1000243697,  425127.499532109]
, [553067.4487768516,  480630.20064410876,  564240.0599785547,  634691.1117343208,  672476.4662604374,  678919.0273557549,  780812.3800457495,  582554.5315090489,  521731.72586360545,  582763.0785912204,  676304.9435025617,  699520.9180088551,  703564.409614038,  818460.8002988051,  578379.2426597459,  470832.9908973022,  643761.9805416813,  660485.4401467816,  722675.8676362758,  701800.5361184613,  864278.0823532302,  542905.6253892162,  459977.7448273293,  662781.2060944121,  664907.1642697613,  727982.3777073738,  699614.7304072687,  876620.961957828,  516321.709027753,  432728.2639609]
, [572328.9030389525,  490593.93031862593,  565106.2572133985,  638657.542698048,  668684.2643453919,  683596.9116216521,  760574.3146353568,  599882.4308863136,  526186.5023143083,  588016.8502845285,  673603.5032407119,  683109.8100053291,  715228.1009897675,  808796.8454219179,  594373.6344964767,  478163.179498207,  645383.7296268827,  656899.1631252733,  718027.8906676488,  704632.1634951205,  854058.1098891377,  554920.8755706808,  477842.5258215334,  651038.6047745846,  672817.5291232929,  720267.5677829974,  709367.6125909211,  866548.31554998,  529015.6035736826,  444519.2478471105]
, [555344.2533522644,  484668.43186292675,  571750.0956801557,  635315.3658417129,  667912.0077526044,  669743.336630828,  753255.9707510415,  588951.3433416081,  526201.865829121,  574863.3507935182,  674941.2826105213,  679821.290819602,  699074.189864433,  816719.0108830844,  586860.8742009194,  475927.5316139096,  643369.9180431777,  654259.5763267188,  702886.1718881702,  698600.2146670307,  843249.2316596024,  564882.8058952163,  480645.0071296823,  654849.6733610227,  659503.5352564532,  715028.3104682665,  689547.9589708735,  858381.9683456195,  541517.4127327997,  462672.21993711486]
, [548011.6000685965,  476264.2588428237,  571600.1915780472,  633276.7868491831,  664736.0423954478,  672973.2733202995,  774680.3493922228,  569308.3480072769,  510558.4103204227,  590505.4066852013,  684588.4313415203,  684253.6337099023,  704151.97391196,  825285.1470153244,  565177.2531828578,  468999.7499754024,  645194.1256031852,  666729.4214096487,  707293.6520569711,  707408.439506766,  862309.1478191038,  544705.8482495412,  457280.1998310754,  667542.6816943667,  663080.345246777,  721769.7751529386,  697083.7043262498,  880731.0781565773,  516618.0253822908,  441723.24329857324]
, [579343.1676272725,  490663.4884913212,  552342.5970229667,  643408.2925819827,  670720.2159503549,  685427.8101524634,  749290.4280183378,  601937.1041650767,  526407.0085477679,  576614.9677864567,  677710.0640695007,  676033.3057162433,  712428.7425259514,  816804.0607791118,  592577.089580348,  492430.64279525675,  630815.3663267234,  672285.9064413926,  702303.6379776709,  712945.6525607477,  838427.7238613942,  569741.4279500949,  484504.7163713506,  648136.2231568391,  671303.8479460153,  713786.9794291969,  702937.0007814813,  864914.251866395,  537395.3235094364,  458610.0076716066]
, [568603.828657702,  475538.03018878226,  577203.3464017743,  661663.8413339149,  639688.928443648,  704269.9219012808,  750800.8965480884,  593216.5351881152,  515647.10177572543,  584184.3019450072,  698430.6348498722,  649118.6574944463,  740049.8666754216,  818196.9671618718,  567703.7013227757,  469755.3121365148,  632261.0835958666,  694784.6605306823,  689933.4175401204,  732499.266986551,  857237.2253457073,  545897.946868791,  466903.62884108553,  647367.5626892883,  685085.0397207112,  701258.6182182076,  714587.0278234476,  868877.9568697064,  517105.9001007572,  445996.69733946974]
, [574716.484604419,  484118.17358334176,  549242.1569658557,  642278.157973054,  672963.9764382057,  684299.9836598729,  744346.2554021153,  603459.9833009696,  518495.17176841764,  573314.0627263102,  679425.1950268486,  687753.6245454281,  712542.5981449465,  797255.2787455319,  601279.5045636916,  481349.32839022507,  633083.2025501813,  668601.8085334775,  717123.3351183822,  707977.418113916,  841553.0234550047,  568308.0308938557,  482409.0538709022,  642047.492829866,  676711.3067969031,  719230.3023934137,  708904.7736548628,  861865.3938589186,  543180.0385382569,  457642.4201525504]
, [573164.356577939,  483227.9439955523,  543990.6011201655,  640622.233728076,  674270.3409216161,  683164.9968729316,  740297.0377607701,  604613.1125757959,  517881.98346948344,  566863.6153360506,  678474.754275942,  688493.502925486,  712307.6533332838,  790716.682824359,  604696.7287453864,  481500.0575105747,  629873.6702593559,  667452.6943716325,  716955.8317960734,  707674.0083589269,  838281.9637266619,  570983.9469427705,  483273.4936432859,  640299.8672526713,  676517.8373970756,  719074.698913267,  708872.740674624,  859987.9589242836,  545273.8472562485,  459126.48458318796]
, [570618.6840034818,  480916.65276923537,  547478.8688370764,  639154.1881419921,  669452.9491201915,  680869.3721939855,  742165.2598012908,  598422.8039925412,  514630.3602162568,  571480.0505751675,  676207.6110391137,  684438.7580083266,  709193.5863419559,  795496.0229270303,  596294.5587641447,  477667.2676679947,  631905.7395733473,  665691.1954264102,  714579.9194499253,  705198.49463121,  840674.721994564,  564335.3169874743,  479411.46326961607,  641832.560934428,  674517.9441280672,  717566.5246377861,  706875.5324762978,  861326.1546107476,  540618.4818508491,  455868.36754136905]
, [532567.7334912753,  468316.4917985418,  579005.3633711151,  652757.2265418682,  680936.4685952791,  680112.6350804174,  770544.9488093997,  562265.1512331621,  514177.81064056663,  589637.3834801143,  665546.5775324184,  703735.9641878214,  703417.412678218,  838143.900200967,  563086.484774579,  445156.4103366907,  664316.541708005,  651774.9699151297,  725194.3109703477,  697373.9866591818,  873736.0258333558,  530712.8140057231,  463382.71897327784,  664477.6840057333,  663427.500380098,  724400.038000822,  700306.3771286121,  873708.0633364252,  524574.5418505494,  429063.47986773954]
, [547578.6479268052,  477009.62975176907,  564690.457960815,  634430.4524260338,  673817.2352162666,  679651.5648802433,  786956.8289900919,  572087.4417154322,  514891.31463111506,  584326.1299878577,  673542.531491097,  704460.7235649123,  701946.2655963015,  818421.4053863684,  570997.8764596685,  470134.93553247047,  639977.6634097705,  663235.8711950345,  723722.4189555929,  703528.0758667336,  863634.4406612206,  543377.879435224,  457972.4340280298,  662148.9673908169,  662320.0848834498,  730070.6839871153,  700090.9248489446,  877785.6684183173,  521073.8324557497,  437111.1538077942]
, [575691.4951800206,  482276.3899298861,  548782.1809205026,  644257.9280827078,  675928.760857903,  683218.6084921743,  743664.7103714866,  601017.5762148818,  522342.5092677441,  566979.321198046,  679015.5732510551,  690969.4406265292,  707554.6433762222,  804692.4353246742,  594053.6264575233,  487975.4696557953,  624144.9845826662,  674226.9228182755,  712397.7625710145,  711589.1116445939,  837503.9055491864,  569572.1310468068,  482086.2549887048,  642648.8375738295,  673076.4365898536,  723774.5637587033,  704354.4879254665,  864923.3573298517,  541620.5674094752,  457697.7795325288]
, [573496.8361492082,  481022.0225527562,  543545.2329563807,  642580.382995658,  676734.6818146758,  681546.1884334298,  741748.3756871971,  601634.3197827344,  520726.7713749276,  559502.218968291,  679661.8440965173,  692311.1727546953,  707270.9066596315,  797365.1481444427,  597358.484535656,  488542.7665026905,  618769.5035418957,  673321.7066182874,  711993.1817416493,  711362.8111904514,  833755.9599819372,  572783.7088653288,  482555.22476927354,  640424.6554258862,  672654.0849390747,  723605.1974846898,  704207.12199477,  862448.2338065748,  544086.4740622574,  459832.5786386382]
, [553766.0858266263,  488267.47286240937,  575921.8207877512,  633489.6023630861,  667943.0268979538,  669628.878467528,  754068.2208313864,  575845.7777814585,  526720.648476044,  583668.9472338989,  670101.8621801501,  680671.9985543337,  707230.6953270952,  815402.9681462235,  596592.8698279691,  471077.09310387156,  660075.8443038393,  639934.0860598702,  711921.6567042964,  689435.9913374059,  848795.9452784993,  561852.2257849622,  493850.9951483097,  649954.4353789653,  661284.3538546858,  703928.5760613811,  697418.6633721062,  846065.3667033773,  555330.3319567319,  466202.44735394657]
, [570238.1059005266,  480251.9467210428,  542818.2707648152,  637563.0836463921,  673982.8260583836,  678937.4038971828,  740393.7537779863,  596160.3775587984,  519228.2634192721,  561709.9337977897,  673345.5062724814,  689488.0988597731,  704117.1625665117,  796061.6886067166,  591988.1850989689,  485229.4394821169,  619092.8191145683,  669103.5412247763,  709698.6406744207,  708412.6471843112,  833528.9661326164,  567859.6756609291,  479338.225943821,  640856.1128532601,  669628.6122651584,  721923.3123241884,  702209.4436465406,  862060.3007876388,  541009.1063759491,  457295.6231772516]
, [570281.7681864804,  478930.53777886624,  542575.8336606601,  639641.2316747975,  673867.9106431169,  678805.0752639514,  740111.2579175534,  597277.1711556538,  517941.4957201646,  559160.6444452048,  675602.6776811603,  689319.6520390469,  703992.0615151384,  796228.0772525702,  592134.834555057,  485056.68411912833,  618519.5490447857,  669794.5813889041,  709504.9610080222,  708457.8818822281,  833446.8078735424,  567968.0322320405,  479301.1510859951,  640758.9848637958,  669769.2135767932,  721910.8886603123,  702167.152771926,  862083.6030922596,  541002.7834203709,  457300.6650711169]
, [567357.5601132794,  485725.611290193,  564946.0162382013,  640794.2487079463,  671446.4784675548,  686325.981018359,  763924.457072111,  590546.8024130457,  519966.0928756448,  588209.1387314446,  671480.3272221461,  686511.2021239717,  716702.86420751,  807765.565137006,  587395.2982619593,  479319.5147418485,  639616.8829335679,  658740.5924953739,  718080.0955191501,  708103.9107362351,  852968.3152120685,  555051.7658382991,  476814.75684637984,  648167.5236289899,  671544.237270898,  721830.2407623769,  711812.8797071897,  865482.0163330929,  533545.9127568357,  448570.5606178581]
, [578670.2830122077,  487980.0962177963,  546959.3796634275,  640395.5191828146,  677194.5548620749,  683531.5984431553,  743156.3528733915,  602464.5894974534,  520619.3943001179,  567102.4122905304,  675608.8696680001,  692126.6893263212,  707985.6478318723,  794769.9720895052,  597971.4499445597,  491129.11906952853,  619898.6022036008,  670234.773066819,  711286.353163863,  710364.5106078325,  831899.5600022619,  572567.495381521,  483253.83521605044,  639448.5575229507,  670229.3379316955,  721473.5887075807,  704231.554547169,  859634.4951650188,  544876.8736412914,  460092.8209378612]
, [565314.9184040272,  484063.1739906845,  541206.918773543,  656816.0357683583,  650842.6336700626,  695601.3902810122,  737955.4447021919,  594810.1656694057,  517206.02567918494,  563419.3588121952,  683953.0753782239,  683038.3133945204,  706864.2395619467,  797636.6169016038,  589628.7693886878,  485444.9411504024,  614846.8734284686,  677386.376244084,  702194.2343263414,  712340.7688014895,  832876.799985817,  565744.9027211132,  481498.4288371758,  636646.3659162885,  677470.1895050532,  714961.2849906416,  706084.802882003,  860453.6743999989,  540946.0810647514,  456451.9327695804]
, [565411.3019276479,  483866.19919640524,  536479.0519603562,  656949.3262902519,  650344.7818416471,  686821.1623727435,  740084.5033616337,  599743.8678022705,  514581.50839072186,  558629.200759403,  694642.6828527086,  665295.6608346096,  711558.8309868182,  789875.0151541368,  586657.0502725489,  485976.1258008578,  611909.5610626327,  680730.5710058274,  704348.9382908705,  714123.6929688874,  828525.4028622337,  563055.8605458251,  480012.7694988834,  624367.9605258186,  672431.0186954141,  714238.3802807002,  702975.5509468982,  849030.230162213,  543470.2566104517,  454797.65372214385]
, [551485.7582168678,  484797.7576037137,  574279.6326426578,  636026.6915228424,  669080.3520246419,  670146.5935092,  753815.5206690746,  581111.2766403493,  524312.6201947408,  577017.8577185513,  670795.8297997364,  681583.8622992719,  696955.064531606,  816769.5647246627,  583235.1274678777,  475237.9974250959,  642717.7935984237,  653800.2329758258,  701407.0247490399,  699153.9757672467,  839320.325593707,  564851.2401035347,  482258.54675381124,  653478.9928000426,  656485.5020526106,  714638.8642739374,  688711.9219616082,  856946.2413155764,  545353.5590061658,  464704.42965656164]
, [543800.740959087,  475509.54615985276,  571579.3117824348,  633217.7312172557,  664444.8802815268,  674066.9296602804,  780511.3852067337,  560947.6054458689,  506062.1224394514,  593401.5607335186,  682542.325615062,  689372.9410209395,  701827.9502885883,  827663.5243257191,  558964.284785285,  469079.1648883046,  642775.4465747839,  668239.0718218744,  708337.8858222204,  708652.4655090517,  863925.5445906341,  544190.0068348584,  457180.4640837844,  666269.1199614621,  661622.2433100648,  721791.1823163619,  698290.4867025592,  882430.4663235209,  522160.241440528,  443972.08369073545]
, [566643.8103123811,  485043.8327747854,  542724.7502254263,  654241.9432998487,  648920.2254429667,  695746.2324564134,  736728.2483183579,  593795.1049699516,  515499.6623619259,  560071.6395777275,  694169.6527447123,  672063.5814263468,  714843.0834584027,  797478.7005929205,  584986.2080567868,  489663.7786695153,  614436.1617584829,  676487.5862319788,  697156.2537271893,  715113.6515861596,  833824.5521163502,  563023.2985367649,  484177.803424303,  636067.5360069614,  676148.1950365033,  711334.3989383805,  707491.9402559362,  861810.921085008,  539355.5320114607,  459057.0091023812]
, [579926.7620858608,  485123.71068965504,  547845.3990345667,  645201.1957319914,  668719.6303454236,  689414.0864762455,  739946.1638577302,  604140.9070397662,  520770.71616674814,  567718.2555875673,  681492.3494992291,  675109.7462489237,  710749.1410546459,  804073.2134998529,  591231.4148364601,  492169.635283215,  618101.1756872075,  676413.7461650525,  699127.8196366736,  714848.9318778883,  833429.7206643636,  568602.1709246014,  485462.9587432937,  634981.5071613655,  676246.1314451551,  712015.5112834938,  707087.5002000371,  862721.7246362765,  541063.2136098866,  460312.08428024367]
, [578973.6346115347,  484975.5525807626,  546216.5597252399,  644339.2574880259,  668593.657594039,  688508.6862763803,  740917.9754246764,  603973.1842099727,  520311.2582451109,  566364.9241810918,  682720.2306732484,  676071.4636036623,  711111.9497615979,  801643.2731256271,  591982.3400177617,  492228.19008447736,  617257.563622328,  676207.0642220067,  698964.0150426937,  714820.1799888111,  833253.9184579268,  569051.7588425167,  485442.96783428796,  634863.5257458093,  676240.4122842271,  711995.4166026994,  707071.9619328587,  862756.2933688996,  541345.5529287225,  460630.2050736169]
, [581324.0488280631,  495888.06723259157,  560549.556262464,  644235.1290228784,  665320.8057303254,  683230.1348256322,  754093.4921877148,  600561.5746570326,  522733.43758622144,  574275.7464309147,  679971.920769406,  679865.9589265705,  704801.5817819312,  805698.4232081188,  592343.7036889682,  499090.1121744162,  623339.9464236603,  676674.9025439459,  701770.4555962154,  709999.6365668622,  838665.1229271136,  569117.7960723898,  483981.28918792435,  641632.2292082543,  671516.006673379,  713466.5594272364,  703489.3022181412,  864929.0285843465,  543030.3500290142,  464069.3270612596]
, [583433.558173317,  491598.9642039728,  547019.9012276541,  641992.5698422418,  670155.0439486479,  693823.6436655385,  722638.9076738971,  608730.9480723124,  525321.863651913,  567441.4752985431,  679060.4565813935,  671322.3894653143,  708763.5733918662,  801145.2826792842,  596549.5232118987,  497245.0004445325,  617640.5847624935,  674330.7156009518,  695149.0934446133,  713611.9398679234,  829171.175378786,  574664.8972062131,  490660.65898529324,  634245.963822113,  673824.2915142181,  711328.9481698611,  701531.2995757124,  861241.9649214356,  546628.7083690463,  466332.89798645565]
, [577578.6242873674,  484500.5652380709,  545087.416002337,  643073.4688859037,  668629.3710074436,  687674.3904827064,  742829.4956762823,  603015.0463514948,  519709.83361716435,  564690.2705786456,  682955.6683841384,  677343.8685743293,  711100.2282758421,  800090.104827746,  592760.0228897054,  492233.1140079197,  614817.7555381042,  675769.3544911833,  699052.2130007566,  714800.985111807,  832137.3966514589,  570910.4267208637,  484977.7048240875,  633661.7445938227,  675535.1912153092,  711985.0694742565,  706769.026145722,  860830.8658553592,  543119.2847041088,  462189.0132749044]
, [575071.7915208398,  484883.52403141634,  544020.1562624733,  638283.4963481687,  672877.1086360579,  687777.8408919995,  744322.9200518052,  599023.061332728,  516200.1303188472,  562049.5341016804,  681685.8717225143,  681954.2746122506,  709910.677328053,  800697.595843038,  592175.1003344437,  490681.5884119326,  615583.9330945202,  677383.0690673703,  704607.8710605282,  714014.5085164482,  833320.264095432,  570490.5824417091,  481371.58014649554,  634624.2833166486,  676243.9554091765,  712002.4512578138,  707068.3740552858,  862653.8013397385,  546037.339384777,  462668.50468258193]
, [579812.2286959301,  490235.4874551235,  538201.3396592907,  640829.0375265306,  671055.0601366473,  690726.3782148446,  715385.2433616836,  607325.6197281255,  523948.40946368413,  562720.190170704,  680756.9380602295,  675855.100249203,  709142.0337748306,  797270.2393490114,  597951.4281944978,  497094.1099779984,  618789.4605063605,  673868.5820237012,  695109.1761740226,  713591.0927534794,  828183.7075600231,  576741.9492414464,  491560.9783063103,  635052.5336876241,  672779.8597717688,  710730.8019576953,  704104.1246875662,  862434.475993629,  550744.2363999332,  469279.5646864849]
, [581603.7489517694,  488144.00120139896,  546389.1754266389,  640648.6524901695,  665195.0060465302,  685420.1494426006,  740075.7448129453,  607462.3768867557,  523667.85180987936,  563977.4264117652,  680925.8975561111,  673912.6537172738,  708968.3523279227,  797000.822025904,  597685.6547694138,  496971.5094103282,  613343.5104429664,  674049.7049426624,  694906.8855007796,  713513.4499362749,  829847.9039894457,  575297.029659812,  490690.0558636297,  633981.1752782167,  671152.571596536,  711598.1732227812,  703419.1386593349,  864554.3604934132,  547698.3202774422,  466189.65646136313]
, [577819.2257207022,  484513.31187654583,  545153.9261121536,  643255.050307633,  668700.5486549989,  687788.4797075705,  742470.0521959774,  602837.0431331261,  519579.6030607115,  564509.7572452627,  683088.4324184946,  677584.300825449,  710969.3275390636,  799767.7658822065,  592746.8233751081,  492231.6196545063,  614565.9283009318,  677812.4440225107,  702952.350155761,  715606.5583817669,  832279.6746707099,  571706.0232465733,  482475.4804495116,  634124.0141523157,  675953.6383566436,  711984.2151183158,  706900.3989184169,  861691.3707785181,  545710.4651454893,  461697.14402737987]
, [577213.7251457034,  488146.6397672072,  546336.0134041935,  640428.5754529781,  665055.2981832423,  685238.3139817254,  740504.4000027983,  606694.1496356124,  523691.49405685195,  564170.3961027024,  683361.1219473283,  678821.9615775306,  708172.6931817618,  794922.5189718473,  597818.3054260339,  497060.4276357671,  614145.0475264875,  673891.4593616679,  695064.717279599,  713503.4489603179,  829158.535895252,  576166.8095125463,  490373.5168173711,  633664.7064092037,  673637.8872990742,  711602.8965921238,  702217.9605810955,  861593.3422818093,  546924.9999356976,  463259.43714283855]
, [579530.3596113469,  489067.78995132266,  538897.332058291,  637938.7553067873,  670798.0199771113,  692108.7407607867,  720606.3226789783,  607905.9804067693,  524818.3200591769,  563561.528113705,  679256.2868438106,  672974.271913701,  708401.2257133863,  799672.2602089257,  597403.3942431206,  496461.74703917105,  616634.6693201695,  675421.9412863003,  694998.5832983888,  715717.772719095,  825130.5901846393,  575998.3621082613,  491397.9438015807,  632483.3680362821,  672771.1788603979,  711675.6644052013,  701144.0675127842,  863138.0308455635,  541784.4578436265,  1.0]
, [567440.0262471181,  471946.5542686427,  582492.3878524891,  664889.4515243642,  632546.9996696521,  712414.4106064951,  741825.3607327823,  580769.8745169822,  511945.914978075,  585487.3175840032,  700955.470144795,  641098.6089066763,  735378.6597771306,  822770.7830884503,  563676.870573938,  472252.64247805317,  628215.1135769851,  689962.9352515909,  669797.5775856473,  734410.7768633884,  858797.9557015456,  543069.5937226714,  472662.3117737677,  640367.2823704836,  681410.4820339688,  692483.4879978287,  719629.3483132265,  881477.7924973662,  522008.9022788477,  449101.11605213897]
, [538168.1704864606,  472646.9887525539,  576275.1857957073,  649670.2049029294,  678714.7487539336,  679636.3268831688,  769005.1723535382,  571932.6855799779,  520745.42201751313,  588625.5241870236,  669734.0812561769,  700165.703601941,  704611.1780985363,  835810.7850795253,  568145.5785770423,  448141.3717125892,  664015.4571051507,  649699.2374926533,  724834.4982983236,  696428.4159407737,  873346.0076267376,  530715.5891441605,  463656.2474834712,  664955.4394294146,  666142.9749923421,  723493.5249249882,  700118.320646305,  873237.9911067003,  520019.39049228095,  425128.8179100345]
, [553064.1764756889,  480626.9963108024,  564242.1592786352,  634694.8214731759,  672477.3432596087,  678919.1466825063,  780810.7053758916,  582554.4534247976,  521729.0881383793,  582761.1943928953,  676308.5623349968,  699517.1306382145,  703566.5565347731,  818462.1608498392,  578380.3084792005,  470825.4783310699,  643766.5650022823,  660484.2149107696,  722675.962937031,  701799.0296658353,  864279.6598635208,  542903.4928393827,  459978.145141107,  662780.8652949936,  664909.2693808326,  727980.8094497889,  699614.2964264418,  876621.2257087363,  516319.4080246008,  432729.2550187169]
, [559027.1065883308,  489409.6402299615,  572836.2509938416,  633790.2951807512,  668922.4421290933,  671938.4237803125,  755505.6678358983,  588333.5004286602,  532596.268299452,  581446.1394065669,  669144.2912573852,  682818.4330142963,  701470.2312724052,  817855.7824250062,  590067.3732911109,  480993.73702115275,  646925.3265966119,  652410.042284705,  706635.0570999895,  700718.9003122271,  846287.7820570146,  568234.8115097436,  486955.68501234846,  657464.4677759341,  659666.6336394006,  717477.5958003501,  694240.3730349002,  859256.2612678689,  548363.5998529604,  467334.1548388279]
, [572324.5203441425,  490587.1390474264,  565103.1155281207,  638662.24921876,  668687.0821443801,  683596.3100228315,  760575.5706204849,  599883.7601197322,  526182.5666025524,  588006.2861675649,  673607.9492899124,  683112.0952389971,  715225.0660223159,  808800.8553045791,  594373.6374872765,  478155.6252632977,  645381.1827591271,  656901.8732633574,  718030.2266073634,  704628.9164476243,  854059.9268951017,  554918.4571082188,  477839.9220542359,  651035.8946375907,  672820.6616944415,  720269.0654181744,  709363.7654270617,  866549.9673882358,  529010.9091059866,  444516.10061127297]
, [558001.9084776294,  487658.3887651104,  574572.5214551644,  638007.6495350252,  670546.1699050539,  672460.160192567,  755610.7830277783,  591373.0136937721,  528866.4664106998,  577865.3671740423,  677231.3513712374,  682469.3886397934,  701608.5618061007,  818948.0841658235,  589893.4493385302,  479306.4172266703,  645989.0045455595,  656980.4024117545,  705426.6554151771,  701247.4087002737,  845398.4217955123,  568150.7043635616,  483980.9928444961,  657502.3707351685,  662239.7109544377,  717529.1888461557,  692288.257712961,  860425.4317929761,  545072.9676867636,  466181.2696713412]
, [556114.7280141719,  494643.7859229374,  571173.3932149088,  639764.2727976231,  665573.4900872408,  671296.2830132394,  752307.8845951124,  592220.1421783429,  528572.1837973911,  582466.0020626088,  679309.931833698,  670174.050078447,  704712.918361871,  708882.4221744386,  587272.2123069837,  481348.78002889745,  645726.9170825023,  658523.6679174477,  697716.1843168872,  705624.2663529918,  844426.7805595109,  718621.4986763108,  483778.827275694,  658691.630417613,  722273.6523645127,  712397.6230694922,  694745.7189126968,  862830.8297811728,  544629.0286158954,  469971.8457561093]
, [575254.8968630387,  491948.82348348515,  563285.785655195,  631260.936387493,  656469.2463437628,  672123.29822464,  757162.408705644,  600226.1521341384,  526619.1061724639,  579674.2442716629,  680945.3364528813,  673242.6766282432,  700638.8821091969,  808600.8885338567,  592658.4326549014,  492238.6770506074,  632690.2025534734,  668564.765544406,  698537.4960581613,  705630.1637342622,  842546.4344996186,  571387.0790521892,  482854.7872691767,  652315.4098976646,  667448.2222273595,  713074.6336324137,  697989.704600813,  864220.9843678614,  542379.7641369879,  465775.63242193655]
, [548787.1997090465,  477115.6914376789,  572532.8763056928,  634204.6581510813,  665686.6216248086,  673917.059309984,  775562.541288632,  570102.4015895824,  511407.61968577304,  591486.6156581596,  685460.1953875009,  685252.944910525,  705200.3221257998,  826317.5983135549,  566123.766703254,  470033.33253797644,  646255.4405299406,  667746.5007829082,  708339.9826084881,  708501.9610694818,  863357.2349816053,  545769.1068226804,  458370.04771436704,  668664.6711446887,  664171.8847102735,  722860.0773890135,  698192.1954605761,  881797.7602042174,  517743.06508087256,  442871.28644658823]
, [582747.8667609834,  498020.15470546973,  563060.9629622428,  642453.7210321665,  666680.7264249253,  682262.3482908648,  756836.9451520232,  605610.0445819027,  525201.9408300878,  575554.5114342963,  681552.6734279589,  677560.9266230594,  708851.3316747462,  808407.8239356804,  598780.1055169562,  496024.90198630525,  627191.614063607,  673598.5229663379,  701553.2204131129,  708729.515338305,  841278.4286597916,  568900.1601758841,  481098.3001805217,  648859.2542829416,  670526.752255039,  714794.9226359685,  702070.687134127,  865570.9717035136,  538207.2007794501,  461751.7817433475]
, [579486.4676286895,  491034.5350502419,  553117.6100208282,  643760.8392417299,  671057.5603930107,  685848.5054154415,  749193.0683024882,  602257.0439402743,  526666.243006327,  576814.7809261967,  677942.7162253951,  676258.2557221162,  712686.2359355068,  816970.3426822531,  592831.1725486983,  492654.7436022635,  630975.6071655164,  672440.733190268,  702485.7516939704,  713157.2878059143,  838571.4844459251,  570014.4486822523,  484738.46362259483,  648276.7051000546,  671458.3705296246,  713929.3596498892,  703106.3586814974,  865024.605336451,  537646.3044020436,  458819.6755846968]
, [569732.672352952,  475593.42281481763,  581456.471613381,  661672.2859485411,  639688.5455850294,  704272.4565820026,  750833.137570908,  593219.2249175528,  515649.8600288895,  584180.6889362595,  698356.144046869,  649111.8059645691,  739946.3991867042,  818203.6122356841,  567746.3646503162,  469777.79024572397,  632277.0612945087,  688114.2976693076,  677295.5491189533,  732548.7744607403,  857234.9834734831,  545918.8453779279,  466908.0010139599,  647344.7155370079,  683788.1661368004,  704114.7557662869,  705253.643600616,  868933.8044136879,  513078.35514523997,  446517.2242739651]]

all_pred = [np.array(x) for x in all_pred]


#Creating a Metrics instance of SARIMA ensemble

sarima_ensemble = Metrics(sarimax_top_test['normal'][0].actual, sum(all_pred)/len(all_pred),_)
sarima_ensemble.model_name('SARIMAX','Ensemble')


sarima_ensemble.fig()


sarima_ensemble.resid_fig()


sarima_ensemble.get_all()

For model ('SARIMAX', 'Ensemble'):

Mean Squared Error (MSE):  1537901238.130723 
Root Mean Squared Error (RMSE):  39216.08392140555 
Mean Absolute Error (MAE):  30831.44837953442 
Mean Absolute Percentage Error (MAPE):  5.143267751746807 
Mean Absolute Scaled Error (MASE):  None


#Training SARIMA model to predict the future, AR=1, I=1, S=7, S(AR)=1, S(I)=1

set_inferred_index(df)
model = SARIMAX(df.new_cases, order=(1, 1, 0), seasonal_order=(1, 1, 0, 7))
model = model.fit(maxiter=200)


#Getting predictions plus upper and lower intervals for them

sf_sarima = model.get_forecast(100).summary_frame()


fig = go.Figure()
fig.add_scatter(x=df.index[-150:], y=df.new_cases[-150:])
fig.add_scatter(x=(list(sf_sarima.index) + list(sf_sarima.index[::-1])),
                y=(list(sf_sarima.mean_ci_lower) + list(sf_sarima.mean_ci_upper[::-1])),
                fill='toself',mode='none', fillcolor='rgba(61, 64, 91, 0.5)')
fig.add_scatter(x=sf_sarima.index, y=sf_sarima['mean'], line_color='#390099')
custom_legend_name(['Historical New Cases','Confidence Interval','Forecast'], fig_in_func=fig)
fig.update_layout(title_text='SARIMA 100-day Forecast with 95% Confidence Intervals')
fig.update_traces(hovertemplate='<b>%{x}</b><br>%{y}<extra></extra>')
fig.show(config=config)


#Training a Holt Winters model with "averaged" value parameters

hw_params = ('add', 'add', True, 0.5, 0.5, 0.5, 0.5, 7)
actual_hw, predicted_hw, model_hw = get_model_pred(hw_params, aic_bic=False, algo_type='hw')
hw_trained = Metrics(actual_hw, predicted_hw, model_hw)
hw_trained.model_name(hw_params,0)


hw_trained.fig()


hw_trained.resid_fig()


hw_trained.get_all(algo_type='hw')

For model ('add', 'add', True, 0.5, 0.5, 0.5, 0.5, 7):

Mean Squared Error (MSE):  629171883.0802559 
Root Mean Squared Error (RMSE):  25083.298887511904 
Mean Absolute Error (MAE):  20177.087661758476 
Mean Absolute Percentage Error (MAPE):  5.214212284769358 
Mean Absolute Scaled Error (MASE):  0.609218003948892


#Creating a grid of parameters for Holt Winters

trend = ['add','mul',None]
seasonal = ['add','mul',None]
damped = [True, False]
phi = [0.25,0.5,0.75,0.99]
alpha = [0.25,0.5,0.75,1]
beta = [0.25,0.5,0.75,1]
gamma = [0.25,0.5,0.75,1]
seasonal_periods = [0,7]

hw_grid_params = []
for t in trend:
    for s in seasonal:
        temp_s = s
        for d in damped:
            temp_d = d
            for p in phi:
                temp_p = p
                for a in alpha:
                    for b in beta:
                        temp_b = b
                        for g in gamma:
                            temp_g = g
                            for sp in seasonal_periods:
                                b = temp_b
                                p = temp_p
                                g = temp_g
                                s = temp_s
                                d = temp_d
                                if not d:
                                    p = 0
                                if not t:
                                    b = 0
                                    d = False
                                    p = 0
                                if (not s or sp == 0):
                                    g = 0
                                    sp = 0
                                    s = None    
                                hw_grid_params.append((t,s,d,p,a,b,g,sp))


#Removing duplicate parameters

hw_grid_params = list(set(hw_grid_params))


#Function to grid search Holt Winters algorithm

def grid_search_hw(param_list, series_type='normal', split_type='val'):
    grid_hw = []
    with warnings.catch_warnings():
        warnings.filterwarnings("ignore")
        for x in tqdm(param_list):
            try:
                actual_hw, predicted_hw, model_hw = get_model_pred(x, aic_bic=False, algo_type='hw',
                                                                   series_type=series_type, split_type=split_type)
                hw_trained = Metrics(actual_hw, predicted_hw, model_hw)
                hw_trained.model_name(x,0)
                if True in np.isnan(hw_trained.predicted):
                    continue
                grid_hw.append(hw_trained)
            except ValueError: #Negative values in series (lr) do not work with multiplicative trend or seasonality
                pass
    return grid_hw


%%time
#Grid searching Holt Winters with 30-day rolling forecast on all three series

hw_trained_grid_normal = grid_search_hw(hw_grid_params)
hw_trained_grid_boxcox = grid_search_hw(hw_grid_params, series_type='boxcox')
hw_trained_grid_lr = grid_search_hw(hw_grid_params, series_type='lr')

100%|██████████████████████████████████████████████████████████████████████████████| 1476/1476 [24:30<00:00,  1.00it/s]
100%|██████████████████████████████████████████████████████████████████████████████| 1476/1476 [22:37<00:00,  1.09it/s]
100%|██████████████████████████████████████████████████████████████████████████████| 1476/1476 [12:00<00:00,  2.05it/s]

Wall time: 59min 8s


#Concatenating all the results into a single dataframe

hw_df1 = pd.DataFrame([(x.model_attribs,x.rmse()) for x in hw_trained_grid_normal if x.rmse()!=0]
                    ).sort_values(1).head(100)
hw_df1['series'] = 'normal'
hw_df2 = pd.DataFrame([(x.model_attribs,x.rmse()) for x in hw_trained_grid_boxcox if x.rmse()!=0]
                     ).sort_values(1).head(100)
hw_df2['series'] = 'boxcox'
hw_df3 = pd.DataFrame([(x.model_attribs,x.rmse()) for x in hw_trained_grid_lr if x.rmse()!=0]).sort_values(1).head(100)
hw_df3['series'] = 'lr'

hw_df = pd.concat([hw_df1, hw_df2, hw_df3])


hw_df


#Assigning dataframe of top models to variable manually so the time-taking computation can be SKIPPED

hw_df = pd.DataFrame([[('mul', 'add', True, 0.99, 0.25, 0.25, 0.25, 7),20712.180332165215, 'boxcox'],
       [('add', 'add', True, 0.99, 0.25, 0.25, 0.25, 7),20894.851577673555, 'boxcox'],
       [('mul', 'add', False, 0, 0.25, 0.25, 0.25, 7),21026.228299640316, 'boxcox'],
       [('add', 'add', True, 0.75, 0.25, 0.5, 0.25, 7), 21086.6947875702,'boxcox'],
       [('mul', 'add', True, 0.75, 0.25, 0.5, 0.25, 7),21115.427827668307, 'boxcox'],
       [('add', 'add', True, 0.75, 0.25, 0.75, 0.25, 7),21198.430465465248, 'boxcox'],
       [('add', 'add', False, 0, 0.25, 0.25, 0.25, 7),21218.279105715104, 'boxcox'],
       [('mul', 'add', True, 0.75, 0.25, 1, 0.25, 7), 21485.77810059913,'boxcox'],
       [('add', 'add', True, 0.75, 0.25, 1, 0.25, 7), 21507.32040751619,'boxcox'],
       [('add', 'add', True, 0.75, 0.5, 0.25, 0.25, 7),21749.72376906739, 'boxcox'],
       [('mul', 'add', True, 0.75, 0.5, 0.25, 0.25, 7),21757.825030366934, 'boxcox'],
       [('add', 'add', True, 0.5, 0.25, 1, 0.25, 7), 21781.887911918544,'boxcox'],
       [('mul', 'add', True, 0.5, 0.25, 1, 0.25, 7), 21800.801562062374,'boxcox'],
       [('mul', 'add', True, 0.99, 0.5, 0.25, 0.25, 7),21831.146405051615, 'boxcox'],
       [('mul', 'mul', True, 0.75, 0.25, 0.5, 0.25, 7),21848.928970832647, 'normal'],
       [('mul', 'mul', True, 0.75, 0.25, 0.5, 0.25, 7),21854.872397017916, 'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.5, 0.25, 7),21855.144393002494, 'normal'],
       [('add', 'mul', True, 0.75, 0.25, 0.5, 0.25, 7),21857.69734678629, 'boxcox'],
       [('add', 'add', True, 0.75, 0.25, 0.25, 0.25, 7),21859.87279676428, 'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.25, 0.25, 7),21868.852940747343, 'normal'],
       [('add', 'add', True, 0.99, 0.5, 0.25, 0.25, 7),21876.565488350803, 'boxcox'],
       [('mul', 'add', False, 0, 0.5, 0.25, 0.25, 7), 21888.033915979,'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.25, 0.25, 7),21901.17790621034, 'boxcox'],
       [('mul', 'add', True, 0.75, 0.25, 0.25, 0.25, 7),21915.888995123492, 'boxcox'],
       [('add', 'add', False, 0, 0.5, 0.25, 0.25, 7), 21959.961101636818,'boxcox'],
       [('mul', 'mul', True, 0.75, 0.25, 0.25, 0.25, 7),21969.946942274095, 'boxcox'],
       [('mul', 'add', True, 0.99, 0.25, 0.5, 0.25, 7),21972.038168795618, 'boxcox'],
       [('mul', 'mul', True, 0.75, 0.25, 0.25, 0.25, 7),22010.562546185924, 'normal'],
       [('add', 'add', True, 0.99, 0.25, 0.5, 0.25, 7),22070.34668039192, 'boxcox'],
       [('add', 'add', True, 0.5, 0.5, 0.25, 0.25, 7), 22082.33072523937,'boxcox'],
       [('mul', 'add', True, 0.5, 0.5, 0.25, 0.25, 7),22087.648712996648, 'boxcox'],
       [('mul', 'add', False, 0, 0.25, 0.5, 0.25, 7), 22146.982771968145,'boxcox'],
       [('add', 'add', True, 0.5, 0.25, 0.75, 0.25, 7),22198.745396344057, 'boxcox'],
       [('add', 'add', True, 0.5, 0.5, 0.5, 0.25, 7), 22240.578969418544,'boxcox'],
       [('add', 'add', False, 0, 0.25, 0.5, 0.25, 7), 22251.715603515644,'boxcox'],
       [('add', 'add', True, 0.25, 0.5, 0.25, 0.25, 7),22260.874550086533, 'boxcox'],
       [('add', 'add', True, 0.25, 0.5, 0.5, 0.25, 7),22267.027395884914, 'boxcox'],
       [('mul', 'mul', True, 0.99, 0.25, 0.25, 0.25, 7),22271.401658998955, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 1, 0.25, 7), 22286.645527798642,'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 1, 0.25, 7), 22291.861739448887,'normal'],
       [('mul', 'mul', True, 0.5, 0.25, 1, 0.25, 7), 22298.36786153884,'boxcox'],
       [('mul', 'mul', True, 0.5, 0.25, 1, 0.25, 7), 22316.22940227011,'normal'],
       [('add', 'add', True, 0.75, 0.5, 0.5, 0.25, 7), 22325.18745149854,'boxcox'],
       [('mul', 'mul', True, 0.99, 0.25, 0.25, 0.25, 7),22335.050437416838, 'normal'],
       [(None, 'add', False, 0, 0.5, 0, 0.25, 7), 22336.90704182524,'boxcox'],
       [('add', 'add', True, 0.25, 0.5, 0.75, 0.25, 7),22340.187621136578, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 0.75, 0.25, 7),22351.993759732097, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 0.75, 0.25, 7),22366.221821733594, 'normal'],
       [('mul', 'mul', True, 0.5, 0.25, 0.75, 0.25, 7),22375.719811004143, 'boxcox'],
       [('mul', 'mul', True, 0.5, 0.25, 0.75, 0.25, 7),22415.16532094281, 'normal'],
       [('mul', 'add', True, 0.25, 0.5, 1, 0.25, 7), 22470.74257099917,'boxcox'],
       [('add', 'add', True, 0.25, 0.5, 1, 0.25, 7), 22470.9246036948,'boxcox'],
       [('mul', 'mul', False, 0, 0.25, 0.25, 0.25, 7), 22497.81588643475,'normal'],
       [('mul', 'mul', True, 0.75, 0.25, 0.75, 0.25, 7),22513.4803454825, 'normal'],
       [('mul', 'mul', True, 0.75, 0.25, 0.75, 0.25, 7),22522.323767144033, 'boxcox'],
       [(None, 'mul', False, 0, 0.5, 0, 0.25, 7), 22549.849723870833,'normal'],
       [(None, 'mul', False, 0, 0.5, 0, 0.25, 7), 22553.83159072885,'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.75, 0.25, 7),22578.00314882439, 'boxcox'],
       [('mul', 'add', True, 0.5, 0.5, 0.75, 0.25, 7), 22605.17687762891,'boxcox'],
       [('add', 'add', True, 0.5, 0.5, 0.75, 0.25, 7),22611.479010298655, 'boxcox'],
       [('mul', 'mul', False, 0, 0.25, 0.25, 0.25, 7), 22616.31897540984,'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.75, 0.25, 7),22628.59582758851, 'normal'],
       [('add', 'mul', True, 0.99, 0.25, 0.25, 0.25, 7),22678.872820747092, 'boxcox'],
       [('add', 'mul', True, 0.25, 0.5, 0.25, 0.25, 7),22684.56113894306, 'normal'],
       [('mul', 'mul', True, 0.25, 0.5, 0.25, 0.25, 7),22690.74526891189, 'normal'],
       [('add', 'mul', True, 0.25, 0.5, 0.25, 0.25, 7),22701.376417213047, 'boxcox'],
       [('mul', 'mul', True, 0.25, 0.5, 0.25, 0.25, 7),22704.374929159265, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 0.5, 0.25, 7),22771.593345491412, 'boxcox'],
       [('mul', 'mul', True, 0.5, 0.25, 0.5, 0.25, 7),22806.332457612687, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.25, 0.5, 0.25, 7),22812.201176934097, 'normal'],
       [('add', 'add', True, 0.75, 0.25, 0.25, 0.5, 7), 22812.236004888,'boxcox'],
       [('mul', 'add', True, 0.99, 0.25, 0.25, 0.5, 7),22819.614255215787, 'boxcox'],
       [('mul', 'add', True, 0.75, 0.25, 0.5, 0.5, 7),22833.996968416628, 'boxcox'],
       [('mul', 'add', True, 0.75, 0.25, 0.25, 0.5, 7),22838.390316054752, 'boxcox'],
       [('add', 'add', True, 0.75, 0.25, 0.5, 0.5, 7),22839.294043828057, 'boxcox'],
       [('mul', 'mul', True, 0.5, 0.25, 0.5, 0.25, 7), 22883.74636489252,'normal'],
       [('add', 'mul', True, 0.25, 0.5, 0.5, 0.25, 7),22919.455075977105, 'normal'],
       [('mul', 'mul', True, 0.25, 0.5, 0.5, 0.25, 7),22926.846083076256, 'normal'],
       [('add', 'mul', True, 0.25, 0.5, 0.5, 0.25, 7),22947.984164245587, 'boxcox'],
       [('mul', 'mul', True, 0.25, 0.5, 0.5, 0.25, 7),22951.522764376943, 'boxcox'],
       [('add', 'add', True, 0.99, 0.25, 0.25, 0.5, 7),22961.838268069292, 'boxcox'],
       [('add', 'add', True, 0.5, 0.25, 0.5, 0.25, 7),22963.287571532648, 'boxcox'],
       [('add', 'mul', True, 0.75, 0.25, 0.25, 0.5, 7),22974.21881902943, 'normal'],
       [('add', 'mul', True, 0.75, 0.25, 0.25, 0.5, 7),22977.726384124595, 'boxcox'],
       [('mul', 'add', True, 0.5, 0.25, 0.5, 0.25, 7), 22996.94864516798,'boxcox'],
       [('add', 'add', True, 0.75, 0.5, 0.75, 0.25, 7),23007.75265635506, 'boxcox'],
       [('mul', 'mul', True, 0.75, 0.25, 0.25, 0.5, 7),23010.484581076187, 'boxcox'],
       [('add', 'mul', True, 0.5, 0.5, 0.25, 0.25, 7), 23013.57922637515,'normal'],
       [('add', 'add', True, 0.5, 0.25, 1, 0.5, 7), 23017.86743686907,'boxcox'],
       [('mul', 'mul', True, 0.5, 0.5, 0.25, 0.25, 7), 23018.15633717269,'normal'],
       [('mul', 'add', True, 0.5, 0.25, 1, 0.5, 7), 23022.601434996974,'boxcox'],
       [('mul', 'add', True, 0.99, 0.25, 0.25, 0.5, 7),23028.001425374503, 'normal'],
       [('mul', 'mul', True, 0.75, 0.25, 0.25, 0.5, 7),23042.20657757218, 'normal'],
       [('add', 'mul', True, 0.5, 0.5, 0.25, 0.25, 7), 23044.09126119537,'boxcox'],
       [('mul', 'mul', True, 0.5, 0.5, 0.25, 0.25, 7), 23046.07329097096,'boxcox'],
       [('mul', 'add', False, 0, 0.25, 0.25, 0.5, 7), 23088.734622922697,'boxcox'],
       [('add', 'add', True, 0.5, 0.25, 0.75, 0.5, 7),23108.610331120308, 'boxcox'],
       [('add', 'mul', False, 0, 0.25, 0.25, 0.25, 7),23117.803302414028, 'boxcox'],
       [('mul', 'add', True, 0.5, 0.25, 0.75, 0.5, 7),23118.020468040006, 'boxcox'],
       [('mul', 'add', True, 0.5, 0.5, 1, 0.25, 7), 23177.406984926987,'boxcox']])
hw_df.columns = [0,1,'series']


#Checking series type of top performing 100 models

hw_df.series.value_counts()

boxcox    76
normal    24
Name: series, dtype: int64


%%time
#Training the top 100 models to predict on the 2nd Validation set

hw_val2_normal = grid_search_hw(hw_df[hw_df['series'] == 'normal'][0].values, split_type='val2', series_type='normal')
hw_val2_boxcox = grid_search_hw(hw_df[hw_df['series'] == 'boxcox'][0].values, split_type='val2', series_type='boxcox')

100%|██████████████████████████████████████████████████████████████████████████████████| 24/24 [00:17<00:00,  1.35it/s]
100%|██████████████████████████████████████████████████████████████████████████████████| 76/76 [01:12<00:00,  1.04it/s]

Wall time: 1min 30s


#Concatenating results to a single dataframe, and adding new features to help select best generalising model

hw_top_100 = pd.concat([pd.DataFrame([(x.model_attribs,x.rmse(),'normal') for x in hw_val2_normal]),
                        pd.DataFrame([(x.model_attribs,x.rmse(),'boxcox') for x in hw_val2_boxcox])])
hw_top_100 = hw_top_100.set_index([2,0]).sort_values([2,0])
hw_top_100['Validation 1'] = hw_df.set_index(['series',0]).sort_values(['series',0])[1]
hw_top_100['Avg'] = (hw_top_100[1]+hw_top_100['Validation 1'])/2
hw_top_100 = hw_top_100.sort_values('Avg')
hw_top_100['Diff'] = abs(hw_top_100[1] - hw_top_100['Validation 1'])
hw_top_100 = hw_top_100.rename(columns={1:'Validation 2'})


pd.set_option('display.float_format', '{:.2f}'.format)
hw_top_diff = hw_top_100.sort_values('Diff').head(10).sort_values('Avg')
hw_top_diff


#Training the best generalising model to see its resulting graphs and metrics

hw_params = ('mul', 'add', False, 0, 0.25, 0.25, 0.5, 7)
actual_hw, predicted_hw, model_hw = get_model_pred(hw_params,aic_bic=False,
                                                   algo_type='hw', series_type='boxcox', split_type='val2')
hw_trained = Metrics(actual_hw, predicted_hw, model_hw)
hw_trained.model_name(hw_params,0)


hw_trained.fig()


hw_trained.resid_fig()


hw_trained.get_all()

For model ('mul', 'add', False, 0, 0.25, 0.25, 0.5, 7):

Mean Squared Error (MSE):  865972602.1837779 
Root Mean Squared Error (RMSE):  29427.41242759509 
Mean Absolute Error (MAE):  23802.43029526544 
Mean Absolute Percentage Error (MAPE):  4.712387143258975 
Mean Absolute Scaled Error (MASE):  None


#Training the best generalising model on the test set. (The selected model did not converge so a neighbouring model
#is picked)

hw_params = ('add', 'mul', False, 0, 0.25, 0.25, 0.5, 7)
actual_hw, predicted_hw, model_hw = get_model_pred(hw_params,aic_bic=False,
                                                   algo_type='hw', series_type='boxcox', split_type='test')
hw_trained = Metrics(actual_hw, predicted_hw, model_hw)
hw_trained.model_name(hw_params,0)


hw_trained.fig()


hw_trained.resid_fig()


hw_trained.get_all()

For model ('add', 'mul', False, 0, 0.25, 0.25, 0.5, 7):

Mean Squared Error (MSE):  870653426.6350911 
Root Mean Squared Error (RMSE):  29506.836947309195 
Mean Absolute Error (MAE):  22361.4580045652 
Mean Absolute Percentage Error (MAPE):  3.693766006616543 
Mean Absolute Scaled Error (MASE):  None


#Training the best models

hw_top_diff = hw_top_diff.reset_index()
hw_test_normal = grid_search_hw(hw_top_diff[hw_top_diff[2] == 'normal'][0].values, split_type='test', series_type='normal')
hw_test_boxcox = grid_search_hw(hw_top_diff[hw_top_diff[2] == 'boxcox'][0].values, split_type='test', series_type='boxcox')

100%|████████████████████████████████████████████████████████████████████████████████████| 3/3 [00:02<00:00,  1.19it/s]
100%|████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:06<00:00,  1.05it/s]


#Adding predictions of best models to a list and creating a Metrics instance for graphs and scores

hw_all_pred = [x.predicted for x in hw_test_normal] + [x.predicted for x in hw_test_boxcox]
hw_ensemble = Metrics(hw_test_normal[0].actual, sum(hw_all_pred)/len(hw_all_pred),_)
hw_ensemble.model_name('Holt Winters','Ensemble')


hw_ensemble.fig()


hw_ensemble.resid_fig()


hw_ensemble.get_all()

For model ('Holt Winters', 'Ensemble'):

Mean Squared Error (MSE):  969493609.1950113 
Root Mean Squared Error (RMSE):  31136.692329067504 
Mean Absolute Error (MAE):  23773.820398569882 
Mean Absolute Percentage Error (MAPE):  3.886220076894499 
Mean Absolute Scaled Error (MASE):  None


#Training the best generalising model to predict the future

model_arg = ('add', 'mul', False, 0, 0.25, 0.25, 0.5, 7)
model = ExponentialSmoothing(df.drop(df[df.new_cases == 0].index).new_cases, seasonal_periods=model_arg[7], seasonal=model_arg[1],
                             trend=model_arg[0], damped=model_arg[2])
model = model.fit(smoothing_level=model_arg[4], smoothing_trend=model_arg[5],
                  smoothing_seasonal=model_arg[6], damping_trend=model_arg[3])

C:\Applications\Anaconda3\lib\site-packages\statsmodels\tsa\holtwinters\model.py:922: ConvergenceWarning:

Optimization failed to converge. Check mle_retvals.


sf_hw = model.forecast(100)


fig = go.Figure()
fig.add_scatter(x=df.index[0:], y=df.new_cases[0:])
fig.add_scatter(x=sf_hw.index, y=sf_hw)
custom_legend_name(['Historical New Cases','Forecast'], fig_in_func=fig)
fig.update_layout(title_text='Holt Winters 100-day Forecast')
fig.update_traces(hovertemplate='<b>%{x}</b><br>%{y}<extra></extra>')
fig.show(config=config)


#Splitting main dataframe into train, val1, val2 and test to reset the change in Holt Winters

train_df, test_df, val_df, val2_df = split_train_test_val(df)


#Changing format of dataframe to suit Prophet's standards

train_df = train_df.reset_index()
train_df.columns = ['ds','y']


#Training a basic Prophet model

m = Prophet(yearly_seasonality=False, daily_seasonality=False).fit(train_df)
future = m.make_future_dataframe(30)
basic_proph = Metrics(val_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
basic_proph.model_name('Prophet',0)


basic_proph.fig()


basic_proph.resid_fig()


basic_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  81275569714.3265 
Root Mean Squared Error (RMSE):  285088.70499254524 
Mean Absolute Error (MAE):  283081.08428380836 
Mean Absolute Percentage Error (MAPE):  76.47188503662592 
Mean Absolute Scaled Error (MASE):  None


#Training Prophet with manually tuned hyperparameters

m = Prophet(seasonality_mode='multiplicative', yearly_seasonality=False, daily_seasonality=False, growth='linear',
           changepoint_prior_scale=0.1, seasonality_prior_scale=5, changepoint_range=0.95).fit(train_df)
future = m.make_future_dataframe(30)
manual_proph = Metrics(val_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
manual_proph.model_name('Prophet',0)


manual_proph.fig()


manual_proph.resid_fig()


manual_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  13649368517.613857 
Root Mean Squared Error (RMSE):  116830.51192909264 
Mean Absolute Error (MAE):  89421.1998346818 
Mean Absolute Percentage Error (MAPE):  24.02576661526784 
Mean Absolute Scaled Error (MASE):  None


#Creating copy of dataframes for rolling forecast so as to not modify the original dataframes
rolling_train = train_df.copy()
rolling_val = val_df.copy().reset_index()
rolling_val.columns = ['ds','y']

#30-day rolling prophet forecast, changepoint_scale=0.1
predicted = []
for x in tqdm(range(1,len(rolling_val)+1)):
    m = Prophet(seasonality_mode='multiplicative', yearly_seasonality=False, daily_seasonality=False, growth='linear',
           changepoint_prior_scale=0.1, seasonality_prior_scale=5, changepoint_range=0.95).fit(rolling_train)
    future = m.make_future_dataframe(1)
    pred = m.predict(future.iloc[-1:])['yhat'].values[0]
    predicted.append(pred)
    rolling_train = rolling_train.append(rolling_val.iloc[x-1])

rolling_proph = Metrics(val_df.new_cases, np.array(predicted),_)
rolling_proph.model_name('Prophet',0)

100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [01:11<00:00,  2.38s/it]


rolling_proph.fig()


rolling_proph.resid_fig()


rolling_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  4665891410.810868 
Root Mean Squared Error (RMSE):  68307.33057886883 
Mean Absolute Error (MAE):  54950.77839049637 
Mean Absolute Percentage Error (MAPE):  14.621140502363266 
Mean Absolute Scaled Error (MASE):  None


#Creating copy of dataframes for rolling forecast so as to not modify the original dataframes
rolling_train = train_df.copy()
rolling_val = val_df.copy().reset_index()
rolling_val.columns = ['ds','y']

#30-day rolling prophet forecast, changepoint_scale=0.5
predicted = []
for x in tqdm(range(1,len(rolling_val)+1)):
    m = Prophet(seasonality_mode='multiplicative', yearly_seasonality=False, daily_seasonality=False, growth='linear',
           changepoint_prior_scale=0.5, seasonality_prior_scale=5, changepoint_range=0.95).fit(rolling_train)
    future = m.make_future_dataframe(1)
    pred = m.predict(future.iloc[-1:])['yhat'].values[0]
    predicted.append(pred)
    rolling_train = rolling_train.append(rolling_val.iloc[x-1])

rolling_proph = Metrics(val_df.new_cases, np.array(predicted),_)
rolling_proph.model_name('Prophet',0)

100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [01:25<00:00,  2.85s/it]


rolling_proph.fig()


rolling_proph.resid_fig()


rolling_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  2432587071.990676 
Root Mean Squared Error (RMSE):  49321.26389287561 
Mean Absolute Error (MAE):  44593.85739974021 
Mean Absolute Percentage Error (MAPE):  11.640178102265303 
Mean Absolute Scaled Error (MASE):  None


#Preparing copies of dataframes (for some ease in plotting) for grid search

grid_val = val_df.copy()
grid_val.reset_index(inplace=True)
grid_val.columns=['ds','y']
grid_train = pd.concat([train_df, grid_val])


#Creating the parameter grid for grid search

param_grid = {  
    'changepoint_prior_scale': [0.001, 0.002, 0.01, 0.1, 0.25, 0.5],
    'seasonality_prior_scale': [0.01, 0.1, 1.0, 5.0, 10.0],
    'seasonality_mode': ['additive','multiplicative'],
    'growth': ['linear'],
    'changepoint_range': [0.1, 0.5, 0.8, 0.95]
}
all_params = [dict(zip(param_grid.keys(), v)) for v in itertools.product(*param_grid.values())]


%%time
#Grid search over all 240 parameters with cross-validation

rmses = []
for params in tqdm(all_params):
    m = Prophet(**params, yearly_seasonality=False, daily_seasonality=False)
    m = m.fit(grid_train)
    df_cv = cross_validation(m, initial=str(grid_train.shape[0]-31) + ' days', horizon='1 day', period='1 day')
    df_p = performance_metrics(df_cv, rolling_window=1)
    rmses.append(df_p['rmse'].values[0])

  0%|                                                                                          | 0/240 [00:00<?, ?it/s]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  0%|▎                                                                             | 1/240 [12:12<48:39:08, 732.84s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  1%|▋                                                                             | 2/240 [15:50<28:25:28, 429.95s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  1%|▉                                                                             | 3/240 [19:21<21:42:04, 329.64s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  2%|█▎                                                                            | 4/240 [22:30<17:59:02, 274.33s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  2%|█▋                                                                            | 5/240 [27:21<18:17:24, 280.19s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  2%|█▉                                                                            | 6/240 [30:14<15:50:51, 243.81s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  3%|██▎                                                                           | 7/240 [33:58<15:21:33, 237.31s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  3%|██▌                                                                           | 8/240 [36:57<14:05:34, 218.68s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00

WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
  4%|██▉                                                                           | 9/240 [44:58<19:18:25, 300.89s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
INFO:prophet:Making 30 forecasts with cutoffs between 2021-05-31 00:00:00 and 2021-06-29 00:00:00


#Creating dataframe of all the model parameters along with their RMSE scores

tuning_results = pd.DataFrame(all_params)
tuning_results['rmse'] = rmses
tuning_results = tuning_results.sort_values('rmse')
tuning_results


#Selecting the top 40 models

top_prophet_models = [list(tuning_results.iloc[:40].iloc[x].values) for x in range(40)]


#Training top performing model from grid-search

m = Prophet(changepoint_prior_scale=0.5, seasonality_prior_scale=1.0, seasonality_mode='multiplicative',
            changepoint_range=0.95, yearly_seasonality=False, daily_seasonality=False).fit(train_df)
future = m.make_future_dataframe(30)
grid_proph = Metrics(val_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
grid_proph.model_name('Prophet',0)


grid_proph.fig()


grid_proph.resid_fig()


grid_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  37574992419.43058 
Root Mean Squared Error (RMSE):  193842.70019639784 
Mean Absolute Error (MAE):  159369.0015039467 
Mean Absolute Percentage Error (MAPE):  42.792386741178724 
Mean Absolute Scaled Error (MASE):  None


#Manually passing top 40 models so above time-taking computation can be SKIPPED

top_prophet_models = [[0.5, 5.0, 'multiplicative', 'linear', 0.95, 68496.51956237167],
 [0.5, 1.0, 'multiplicative', 'linear', 0.95, 68536.11617835006],
 [0.5, 0.1, 'multiplicative', 'linear', 0.95, 68628.25718529584],
 [0.5, 10.0, 'multiplicative', 'linear', 0.95, 68942.84671194918],
 [0.5, 0.01, 'multiplicative', 'linear', 0.95, 69628.06674852669],
 [0.25, 1.0, 'multiplicative', 'linear', 0.95, 78758.12814339087],
 [0.25, 10.0, 'multiplicative', 'linear', 0.95, 78759.54733564568],
 [0.25, 0.1, 'multiplicative', 'linear', 0.95, 78924.09412838506],
 [0.25, 5.0, 'multiplicative', 'linear', 0.95, 79151.50283379994],
 [0.25, 0.01, 'multiplicative', 'linear', 0.95, 80213.47033026148],
 [0.5, 5.0, 'additive', 'linear', 0.95, 85405.79814091229],
 [0.5, 10.0, 'additive', 'linear', 0.95, 85422.6856861233],
 [0.5, 0.1, 'additive', 'linear', 0.95, 85459.28342252369],
 [0.5, 1.0, 'additive', 'linear', 0.95, 85463.78934329028],
 [0.5, 0.01, 'additive', 'linear', 0.95, 86569.83984379865],
 [0.25, 5.0, 'additive', 'linear', 0.95, 96911.21875227928],
 [0.25, 1.0, 'additive', 'linear', 0.95, 97216.09695754146],
 [0.25, 0.1, 'additive', 'linear', 0.95, 97405.95622789257],
 [0.25, 10.0, 'additive', 'linear', 0.95, 98005.19271723731],
 [0.25, 0.01, 'additive', 'linear', 0.95, 98697.63093980013],
 [0.1, 5.0, 'multiplicative', 'linear', 0.95, 109049.90195165592],
 [0.1, 0.1, 'multiplicative', 'linear', 0.95, 109274.81993574959],
 [0.1, 1.0, 'multiplicative', 'linear', 0.95, 109998.05918405706],
 [0.1, 10.0, 'multiplicative', 'linear', 0.95, 110067.23623292665],
 [0.1, 0.01, 'multiplicative', 'linear', 0.95, 120940.54246312709],
 [0.001, 5.0, 'multiplicative', 'linear', 0.95, 134317.58224694864],
 [0.001, 1.0, 'multiplicative', 'linear', 0.95, 134429.61859325954],
 [0.001, 10.0, 'multiplicative', 'linear', 0.95, 135015.32116229937],
 [0.5, 0.1, 'multiplicative', 'linear', 0.5, 135445.16340580367],
 [0.5, 1.0, 'multiplicative', 'linear', 0.5, 135474.37422890184],
 [0.5, 10.0, 'multiplicative', 'linear', 0.5, 135482.70052235207],
 [0.5, 5.0, 'multiplicative', 'linear', 0.5, 135545.00735833478],
 [0.25, 0.1, 'multiplicative', 'linear', 0.5, 137338.0141640413],
 [0.001, 0.1, 'multiplicative', 'linear', 0.95, 137342.49488270315],
 [0.25, 1.0, 'multiplicative', 'linear', 0.5, 137377.5775584008],
 [0.25, 5.0, 'multiplicative', 'linear', 0.5, 137383.00742894819],
 [0.25, 10.0, 'multiplicative', 'linear', 0.5, 137388.8692865301],
 [0.5, 1.0, 'additive', 'linear', 0.5, 138373.30412958527],
 [0.5, 10.0, 'additive', 'linear', 0.5, 138424.10538792968],
 [0.5, 5.0, 'additive', 'linear', 0.5, 138440.86432930426]]


%%time
#Finding RMSEs of top models with cross validation, includes 2nd Validation set

grid_val2 = val2_df.copy()
grid_val2.reset_index(inplace=True)
grid_val2.columns=['ds','y']
grid2_train = pd.concat([train_df, grid_val, grid_val2])

rmses2 = []
for x in tqdm(top_prophet_models):
    param_grid = {'changepoint_prior_scale': x[0],
                  'seasonality_prior_scale': x[1],
                  'seasonality_mode': x[2],
                  'growth': x[3],
                  'changepoint_range': x[4]}
    m = Prophet(**param_grid, yearly_seasonality=False, daily_seasonality=False)
    m = m.fit(grid2_train)
    df_cv = cross_validation(m, initial=str(grid2_train.shape[0]-31) + ' days', horizon='1 day', period='1 day')
    df_p = performance_metrics(df_cv, rolling_window=1)
    rmses2.append(df_p['rmse'].values[0])

  0%|                                                                                           | 0/40 [00:00<?, ?it/s]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

  2%|██                                                                              | 1/40 [01:45<1:08:26, 105.30s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

  5%|████▏                                                                              | 2/40 [03:12<59:58, 94.71s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

  8%|██████▏                                                                            | 3/40 [04:43<57:13, 92.78s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 10%|████████▎                                                                          | 4/40 [06:09<54:14, 90.41s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 12%|██████████▍                                                                        | 5/40 [07:37<52:10, 89.43s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 15%|████████████▍                                                                      | 6/40 [08:57<48:53, 86.27s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 18%|██████████████▌                                                                    | 7/40 [10:16<46:11, 83.98s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 20%|████████████████▌                                                                  | 8/40 [11:40<44:40, 83.77s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00

 22%|██████████████████▋                                                                | 9/40 [12:59<42:34, 82.42s/it]INFO:prophet:Making 30 forecasts with cutoffs between 2021-06-30 00:00:00 and 2021-07-29 00:00:00


rmses2 = [49379.81149042711, 49468.19180461303, 49542.06937470975, 49356.75162882476, 50961.68146601285,
          54408.77247706884, 54522.53334574541, 54170.295340920675, 54642.48673508921, 56155.482973485865,
          57632.711162021194, 57592.37418301971, 57443.61971524401, 57775.42374426384, 58843.47959652853,
          68296.25306466538, 68061.10029511034, 68273.48335750909, 67511.87185612555, 69506.19201838046,
          87664.34053715246, 87542.98816953565, 88989.14346179654, 88198.10355742653, 121636.72874178934,
          195005.0039692153, 196573.72459879064, 194608.13077828434, 107505.68359457407, 107882.57339696081,
          107568.7740900801, 107568.27354584601, 112699.90101143483, 195438.89750312644, 112177.88753643607,
          112299.9525291053, 112508.85148867678, 108691.73639965984, 108733.77022216775, 108597.22876546648]


#Finding RMSEs of top models manually, includes 1st Validation set

rmses_manual = []
for x in tqdm(top_prophet_models):
    param_grid = {'changepoint_prior_scale': x[0],
                  'seasonality_prior_scale': x[1],
                  'seasonality_mode': x[2],
                  'growth': x[3],
                  'changepoint_range': x[4]}
    m = Prophet(**param_grid, yearly_seasonality=False, daily_seasonality=False).fit(grid_train)
    future = m.make_future_dataframe(30)
    rmse_manual = Metrics(val_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
    rmse_manual.model_name('Prophet',0)
    rmses_manual.append(rmse_manual)

 62%|███████████████████████████████████████████████████▎                              | 25/40 [01:05<00:35,  2.38s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
 65%|█████████████████████████████████████████████████████▎                            | 26/40 [01:11<00:46,  3.35s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
 68%|███████████████████████████████████████████████████████▎                          | 27/40 [01:17<00:53,  4.11s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
 82%|███████████████████████████████████████████████████████████████████▋              | 33/40 [01:31<00:16,  2.39s/it]WARNING:prophet.models:Optimization terminated abnormally. Falling back to Newton.
100%|██████████████████████████████████████████████████████████████████████████████████| 40/40 [01:50<00:00,  2.77s/it]


#Finding RMSEs of top models manually, includes 2nd Validation set

rmses2_manual = []
for x in tqdm(top_prophet_models):
    param_grid = {'changepoint_prior_scale': x[0],
                  'seasonality_prior_scale': x[1],
                  'seasonality_mode': x[2],
                  'growth': x[3],
                  'changepoint_range': x[4]}
    m = Prophet(**param_grid, yearly_seasonality=False, daily_seasonality=False).fit(grid2_train)
    future = m.make_future_dataframe(30)
    rmse2_manual = Metrics(val2_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
    rmse2_manual.model_name('Prophet',0)
    rmses2_manual.append(rmse2_manual)

100%|██████████████████████████████████████████████████████████████████████████████████| 40/40 [01:45<00:00,  2.64s/it]


#Concatenating all scores into a dataframe, creating new features to judge and find the best model

top_prophet_df = pd.DataFrame(top_prophet_models)
top_prophet_df.rename(columns={5:'rmse_crossval'}, inplace=True)
top_prophet_df['rmse2_crossval'] = rmses2
top_prophet_df['rmse_manual'] = [x.rmse() for x in rmses_manual]
top_prophet_df['rmse2_manual'] = [x.rmse() for x in rmses2_manual]
top_prophet_df['crossval_avg'] = (top_prophet_df['rmse_crossval'] + top_prophet_df['rmse2_crossval'])/2
top_prophet_df['crossval_diff'] = np.abs(top_prophet_df['rmse_crossval'] - top_prophet_df['rmse2_crossval'])
top_prophet_df['manual_avg'] = (top_prophet_df['rmse_manual'] + top_prophet_df['rmse2_manual'])/2
top_prophet_df['manual_diff'] = np.abs(top_prophet_df['rmse_manual']-top_prophet_df['rmse2_manual'])
top_prophet_df['diff-diff'] = np.abs(top_prophet_df['crossval_diff'] + top_prophet_df['manual_diff'])


top_prophet_df.sort_values('diff-diff')


#Training the best model till now to predict on Test set

m = Prophet(changepoint_prior_scale=0.5, seasonality_prior_scale=5, seasonality_mode='multiplicative',
            changepoint_range=0.5, yearly_seasonality=False, daily_seasonality=False).fit(grid_val2)
future = m.make_future_dataframe(30)
test_proph = Metrics(test_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
test_proph.model_name('Prophet',0)

INFO:prophet:n_changepoints greater than number of observations. Using 14.


test_proph.fig()


test_proph.resid_fig()


test_proph.get_all()

For model Prophet:

Mean Squared Error (MSE):  9094464570.419357 
Root Mean Squared Error (RMSE):  95364.90219372827 
Mean Absolute Error (MAE):  76681.47796686129 
Mean Absolute Percentage Error (MAPE):  13.262542335369481 
Mean Absolute Scaled Error (MASE):  None


#Creating an ensemble list of models

prophet_ensemble = (list(top_prophet_df.sort_values('crossval_diff')[[0,1,2,3,4]].head(10).values) +
                    list(top_prophet_df.sort_values('manual_diff')[[0,1,2,3,4]].head(10).values) +
                    list(top_prophet_df.sort_values('diff-diff')[[0,1,2,3,4]].head(10).values))
prophet_ensemble = [tuple(x) for x in prophet_ensemble]
prophet_ensemble= set(prophet_ensemble)


%%time
#Training the models in ensemble list and predicting on the Test set

ensembles_test = []
for x in tqdm(prophet_ensemble):
    param_grid = {'changepoint_prior_scale': x[0],
                  'seasonality_prior_scale': x[1],
                  'seasonality_mode': x[2],
                  'growth': x[3],
                  'changepoint_range': x[4]}
    m = Prophet(**param_grid, yearly_seasonality=False, daily_seasonality=False).fit(grid_val2)
    future = m.make_future_dataframe(30)
    ensemble_test = Metrics(test_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
    ensemble_test.model_name('Prophet',0)
    ensembles_test.append(ensemble_test)

100%|██████████████████████████████████████████████████████████████████████████████████| 15/15 [00:31<00:00,  2.12s/it]

Wall time: 31.8 s


#Creating Metrics instance of predictions from the ensemble of models

ensemble_model = Metrics(test_df.new_cases, sum([x.predicted for x in ensembles_test])/len(ensembles_test),_)
ensemble_model.model_name('Prophet',0)


ensemble_model.fig()


ensemble_model.resid_fig()


ensemble_model.get_all()

For model Prophet:

Mean Squared Error (MSE):  43965126215.34753 
Root Mean Squared Error (RMSE):  209678.62603362204 
Mean Absolute Error (MAE):  178199.43891570877 
Mean Absolute Percentage Error (MAPE):  30.713917699185156 
Mean Absolute Scaled Error (MASE):  None


#Creating copy of main dataframe to format it according to Prophet's standards

proph_df = df.reset_index()
proph_df.columns=['ds','y']


#Training the model and predicting the future

m = Prophet(changepoint_prior_scale=0.5, seasonality_prior_scale=5, seasonality_mode='multiplicative',
            changepoint_range=0.5, yearly_seasonality=False, daily_seasonality=False).fit(proph_df)
future = m.make_future_dataframe(100)
sf_prophet = m.predict(future)
sf_prophet = sf_prophet.iloc[-100:]


fig = go.Figure()
fig.add_scatter(x=proph_df.ds[-150:], y=proph_df.y[-150:])
fig.add_scatter(x=(list(sf_prophet.ds) + list(sf_prophet.ds[::-1])),
                y=(list(sf_prophet.yhat_lower) + list(sf_prophet.yhat_upper[::-1])),
                fill='toself',mode='none', fillcolor='rgba(61, 64, 91, 0.5)')
fig.add_scatter(x=sf_prophet.ds, y=sf_prophet.yhat, line_color='#390099')
custom_legend_name(['Historical New Cases','Confidence Interval','Forecast'], fig_in_func=fig)
fig.update_layout(title_text='Prophet 100-day Forecast with 80% Confidence Intervals')
fig.update_traces(hovertemplate='<b>%{x}</b><br>%{y}<extra></extra>')
fig.show(config=config)


#Resetting dataframe into original splits

train_df, test_df, val_df, val2_df = split_train_test_val(df)


#Function to plot future predictions of all models

def plot_all_models_future(start, future_days, ensemble=False, prophet=True):
    #SARIMA
    set_inferred_index(df)
    model = SARIMAX(df.new_cases, order=(1, 1, 0), seasonal_order=(1, 1, 0, 7))
    model = model.fit(maxiter=200)
    sf_sarima = model.get_forecast(future_days).summary_frame()

    #Holt Winters Exponential Smoothing
    model_arg = ('add', 'mul', False, 0, 0.25, 0.25, 0.5, 7)
    with warnings.catch_warnings():
        warnings.filterwarnings("ignore")
        model = ExponentialSmoothing(df.drop(df[df.new_cases == 0].index).new_cases, seasonal_periods=model_arg[7],
                                     seasonal=model_arg[1], trend=model_arg[0], damped=model_arg[2])
        model = model.fit(smoothing_level=model_arg[4], smoothing_trend=model_arg[5],
                          smoothing_seasonal=model_arg[6], damping_trend=model_arg[3])
        sf_hw = model.forecast(future_days)

    #Prophet
    m = Prophet(changepoint_prior_scale=0.5, seasonality_prior_scale=5, seasonality_mode='multiplicative',
            changepoint_range=0.5, yearly_seasonality=False, daily_seasonality=False).fit(proph_df)
    future = m.make_future_dataframe(future_days)
    sf_prophet = m.predict(future)
    sf_prophet = sf_prophet.iloc[-future_days:]
    
    #Figure
    fig = go.Figure()
    fig.add_scatter(x=df.index[start:], y=df.new_cases[start:])
    if ensemble:
        fig.add_scatter(x=sf_prophet.ds, y=(sf_hw + sf_sarima['mean'])/2)
        custom_legend_name(['Historical New Cases','SARIMA + Holt Winters Forecast'], fig_in_func=fig)
        fig.update_layout(title_text=str(future_days)+' Days Ensemble Forecast')
    else:
        fig.add_scatter(x=sf_prophet.ds, y=sf_sarima['mean'])
        fig.add_scatter(x=sf_prophet.ds, y=sf_hw)
        if prophet:
            fig.add_scatter(x=sf_prophet.ds, y=sf_prophet.yhat)
            custom_legend_name(['Historical New Cases','SARIMA Forecast','Holt Winters Forecast','Prophet Forecast'],
                           fig_in_func=fig)
            fig.update_layout(title_text=str(future_days)+' Days Forecast')
        else:
            custom_legend_name(['Historical New Cases','SARIMA Forecast','Holt Winters Forecast'], fig_in_func=fig)
            fig.update_layout(title_text=str(future_days)+' Days Forecast Without Prophet')
    fig.update_traces(hovertemplate='<b>%{y}</b><br>%{x}<extra></extra>')
    fig.show(config=config)


plot_all_models_future(0,1000)


plot_all_models_future(0,1000,prophet=False)


#Adding Metrics instance of each model to a list

ensemble_all = []

#SARIMA
test_final = train_and_pred(((1, 1, 0), (1, 1, 0, 7)), aic_bic=False, series_type='boxcox', split_type='test')
ensemble_all.append(test_final)

#Holt Winters
hw_params = ('add', 'mul', False, 0, 0.25, 0.25, 0.5, 7)
actual_hw, predicted_hw, model_hw = get_model_pred(hw_params,aic_bic=False,
                                                   algo_type='hw', series_type='boxcox', split_type='test')
hw_trained = Metrics(actual_hw, predicted_hw, model_hw)
hw_trained.model_name(hw_params,0)
ensemble_all.append(hw_trained)

#Prophet
m = Prophet(changepoint_prior_scale=0.5, seasonality_prior_scale=5, seasonality_mode='multiplicative',
            changepoint_range=0.5, yearly_seasonality=False, daily_seasonality=False).fit(grid_val2)
future = m.make_future_dataframe(30)
test_proph = Metrics(test_df.new_cases, np.array(m.predict(future).iloc[-30:]['yhat']), _)
test_proph.model_name('Prophet',0)
ensemble_all.append(test_proph)

INFO:prophet:n_changepoints greater than number of observations. Using 14.


#Creating Metrics instance of ensemble for graphs and score

allmet = Metrics(test_df.new_cases, sum([x.predicted for x in ensemble_all])/3,_)
allmet.model_name('Ensemble of SARIMA, HW & Prophet',0)


allmet.fig()


allmet.resid_fig()


allmet.get_all()

For model Ensemble of SARIMA, HW & Prophet:

Mean Squared Error (MSE):  1976816862.5383713 
Root Mean Squared Error (RMSE):  44461.40868819128 
Mean Absolute Error (MAE):  38336.0939533894 
Mean Absolute Percentage Error (MAPE):  6.5724905529959745 
Mean Absolute Scaled Error (MASE):  None


#Creating Metrics instance of ensemble of SARIMA and HW

allmet_without_proph = Metrics(test_df.new_cases, sum([x.predicted for x in ensemble_all[:2]])/2,_)
allmet_without_proph.model_name('Ensemble of SARIMA and HW',0)


allmet_without_proph.fig()


allmet_without_proph.resid_fig()


allmet_without_proph.get_all()

For model Ensemble of SARIMA and HW:

Mean Squared Error (MSE):  893486223.077247 
Root Mean Squared Error (RMSE):  29891.23990531753 
Mean Absolute Error (MAE):  22716.817463165273 
Mean Absolute Percentage Error (MAPE):  3.7570614363433417 
Mean Absolute Scaled Error (MASE):  None


plot_all_models_future(0,365,ensemble=True)

	Province/State	Country/Region	Lat	Long	...
0	NaN	Afghanistan	33.93911	67.709953	...
1	NaN	Albania	41.15330	20.168300	...
2	NaN	Algeria	28.03390	1.659600	...
3	NaN	Andorra	42.50630	1.521800	...
4	NaN	Angola	-11.20270	17.873900	...

	Date	Country	Latitude	Longitude
47124	2020-01-22	Afghanistan	33.93911	67.709953
47125	2020-01-23	Afghanistan	33.93911	67.709953
47126	2020-01-24	Afghanistan	33.93911	67.709953
47127	2020-01-25	Afghanistan	33.93911	67.709953
47128	2020-01-26	Afghanistan	33.93911	67.709953

	Entity	Code	Year	Literacy rates (World Bank, CIA World Factbook, and other sources)
0	Afghanistan	AFG	1979	18.157681
1	Afghanistan	AFG	2011	31.741119
2	Afghanistan	AFG	2015	38.168041
3	Albania	ALB	2001	98.712982
4	Albania	ALB	2008	95.938637

	iso	country	Democracy index	year
0	afg	Afghanistan	30.6	2006.0
1	afg	Afghanistan	30.4	2007.0
2	afg	Afghanistan	30.2	2008.0
3	afg	Afghanistan	27.5	2009.0
4	afg	Afghanistan	24.8	2010.0

		Validation 2	Validation 1	Avg	Diff
2	0
boxcox	(mul, add, True, 0.99, 0.25, 0.25, 0.5, 7)	29444.34	22819.61	26131.98	6624.73
	(add, add, True, 0.99, 0.25, 0.25, 0.5, 7)	29418.51	22961.84	26190.17	6456.67
	(mul, add, False, 0, 0.25, 0.25, 0.5, 7)	29427.41	23088.73	26258.07	6338.68
normal	(mul, mul, True, 0.75, 0.25, 0.25, 0.5, 7)	30564.44	23042.21	26803.32	7522.23
normal	(add, mul, True, 0.75, 0.25, 0.25, 0.5, 7)	30709.67	22974.22	26841.94	7735.45
boxcox	(mul, mul, True, 0.75, 0.25, 0.25, 0.5, 7)	30785.54	23010.48	26898.01	7775.05
	(add, mul, True, 0.75, 0.25, 0.25, 0.5, 7)	30866.41	22977.73	26922.07	7888.69
	(mul, add, True, 0.75, 0.25, 0.5, 0.5, 7)	31058.85	22834.00	26946.42	8224.85
	(add, add, True, 0.75, 0.25, 0.5, 0.5, 7)	31090.18	22839.29	26964.74	8250.89
normal	(mul, add, True, 0.99, 0.25, 0.25, 0.5, 7)	31001.74	23028.00	27014.87	7973.73

	iso_code	continent	location	date	total_cases	new_cases	new_cases_smoothed	total_deaths	new_deaths	new_deaths_smoothed	...	extreme_poverty	cardiovasc_death_rate	diabetes_prevalence	female_smokers	male_smokers	handwashing_facilities	hospital_beds_per_thousand	life_expectancy	human_development_index	excess_mortality
0	AFG	Asia	Afghanistan	2020-02-24	1.0	1.0	NaN	NaN	NaN	NaN	...	NaN	597.029	9.59	NaN	NaN	37.746	0.5	64.83	0.511	NaN
1	AFG	Asia	Afghanistan	2020-02-25	1.0	0.0	NaN	NaN	NaN	NaN	...	NaN	597.029	9.59	NaN	NaN	37.746	0.5	64.83	0.511	NaN
2	AFG	Asia	Afghanistan	2020-02-26	1.0	0.0	NaN	NaN	NaN	NaN	...	NaN	597.029	9.59	NaN	NaN	37.746	0.5	64.83	0.511	NaN
3	AFG	Asia	Afghanistan	2020-02-27	1.0	0.0	NaN	NaN	NaN	NaN	...	NaN	597.029	9.59	NaN	NaN	37.746	0.5	64.83	0.511	NaN
4	AFG	Asia	Afghanistan	2020-02-28	1.0	0.0	NaN	NaN	NaN	NaN	...	NaN	597.029	9.59	NaN	NaN	37.746	0.5	64.83	0.511	NaN

	iso_code	continent	location	date	total_cases	new_cases	new_cases_smoothed	total_deaths	new_deaths	new_deaths_smoothed	...	extreme_poverty	cardiovasc_death_rate	diabetes_prevalence	female_smokers	male_smokers	handwashing_facilities	hospital_beds_per_thousand	life_expectancy	human_development_index	excess_mortality
75721	OWID_CYN	Asia	Northern Cyprus	2021-01-14	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75722	OWID_CYN	Asia	Northern Cyprus	2021-01-15	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75723	OWID_CYN	Asia	Northern Cyprus	2021-01-16	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75724	OWID_CYN	Asia	Northern Cyprus	2021-01-17	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75725	OWID_CYN	Asia	Northern Cyprus	2021-01-18	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
75930	OWID_CYN	Asia	Northern Cyprus	2021-08-11	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75931	OWID_CYN	Asia	Northern Cyprus	2021-08-12	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75932	OWID_CYN	Asia	Northern Cyprus	2021-08-13	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75933	OWID_CYN	Asia	Northern Cyprus	2021-08-14	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75934	OWID_CYN	Asia	Northern Cyprus	2021-08-15	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

	CountryName	CountryCode	RegionName	RegionCode	Jurisdiction	Date	C1_Flag	C2_Flag	...
0	Aruba	ABW	NaN	NaN	NAT_TOTAL	20200101	NaN	NaN	...
1	Aruba	ABW	NaN	NaN	NAT_TOTAL	20200102	NaN	NaN	...
2	Aruba	ABW	NaN	NaN	NAT_TOTAL	20200103	NaN	NaN	...
3	Aruba	ABW	NaN	NaN	NAT_TOTAL	20200104	NaN	NaN	...
4	Aruba	ABW	NaN	NaN	NAT_TOTAL	20200105	NaN	NaN	...

	CountryCode	Date	...
0	ABW	2020-01-01	...
1	ABW	2020-01-02	...
2	ABW	2020-01-03	...
3	ABW	2020-01-04	...
4	ABW	2020-01-05	...

	country_region_code	country_region	sub_region_1	sub_region_2	metro_area	iso_3166_2_code	census_fips_code	place_id	date	retail_and_recreation_percent_change_from_baseline	grocery_and_pharmacy_percent_change_from_baseline	parks_percent_change_from_baseline	transit_stations_percent_change_from_baseline	workplaces_percent_change_from_baseline	residential_percent_change_from_baseline
0	AE	United Arab Emirates	NaN	NaN	NaN	NaN	NaN	ChIJvRKrsd9IXj4RpwoIwFYv0zM	2020-02-15	0.0	4.0	5.0	0.0	2.0	1.0
1	AE	United Arab Emirates	NaN	NaN	NaN	NaN	NaN	ChIJvRKrsd9IXj4RpwoIwFYv0zM	2020-02-16	1.0	4.0	4.0	1.0	2.0	1.0
2	AE	United Arab Emirates	NaN	NaN	NaN	NaN	NaN	ChIJvRKrsd9IXj4RpwoIwFYv0zM	2020-02-17	-1.0	1.0	5.0	1.0	2.0	1.0
3	AE	United Arab Emirates	NaN	NaN	NaN	NaN	NaN	ChIJvRKrsd9IXj4RpwoIwFYv0zM	2020-02-18	-2.0	1.0	5.0	0.0	2.0	1.0
4	AE	United Arab Emirates	NaN	NaN	NaN	NaN	NaN	ChIJvRKrsd9IXj4RpwoIwFYv0zM	2020-02-19	-2.0	0.0	4.0	-1.0	2.0	1.0

	iso_code	date	country	continent	total_cases	new_cases	new_cases_smoothed	total_deaths	new_deaths	new_deaths_smoothed	...	ContainmentHealthIndex	EconomicSupportIndex	gm_retail_and_recreation	gm_grocery_and_pharmacy	gm_parks	gm_transit_stations	gm_workplaces	gm_residential	literacy_rate	Democracy index
0	ABW	2021-03-29	Aruba	North America	NaN	NaN	NaN	NaN	NaN	NaN	...	65.95	12.5	-8.0	-1.0	51.0	-47.0	-18.0	-3.0	97.5242	NaN
1	ABW	2021-03-30	Aruba	North America	NaN	NaN	NaN	NaN	NaN	NaN	...	65.95	12.5	-4.0	-2.0	46.0	-54.0	-18.0	-4.0	97.5242	NaN
2	ABW	2021-03-31	Aruba	North America	NaN	NaN	NaN	NaN	NaN	NaN	...	65.95	12.5	-1.0	6.0	34.0	-45.0	-18.0	-5.0	97.5242	NaN
3	ABW	2021-04-01	Aruba	North America	NaN	NaN	NaN	NaN	NaN	NaN	...	65.95	62.5	2.0	30.0	37.0	-47.0	-24.0	-5.0	97.5242	NaN
4	ABW	2021-04-02	Aruba	North America	NaN	NaN	NaN	NaN	NaN	NaN	...	65.95	12.5	-50.0	-78.0	35.0	-61.0	-67.0	16.0	97.5242	NaN

	0	1	2
27	((5, 0, 1), (1, 1, 0, 7))	2.866758e+04	boxcox
25	((1, 1, 0), (1, 1, 0, 7))	2.967555e+04	boxcox
70	((5, 0, 1), (1, 1, 0, 7))	2.977680e+04	lr
6	((5, 0, 1), (1, 1, 0, 7))	3.016101e+04	normal
24	((1, 1, 0), (0, 1, 0, 7))	3.093080e+04	boxcox
...	...	...	...
9	((0, 1, 0), (2, 0, 1, 7))	4.350548e+04	normal
26	((0, 1, 0), (1, 0, 5, 7))	4.378069e+04	boxcox
31	((0, 1, 0), (2, 0, 1, 7))	4.409401e+04	boxcox
23	((0, 1, 0), (2, 0, 1, 7))	4.409401e+04	boxcox
1	((1, 0, 0), (1, 0, 0, 7))	6.774864e+10	normal

	0	1	series
895	(mul, mul, True, 0.75, 0.25, 0.5, 0.25, 7)	21848.93	normal
381	(add, mul, True, 0.75, 0.25, 0.5, 0.25, 7)	21855.14	normal
378	(add, mul, True, 0.75, 0.25, 0.25, 0.25, 7)	21868.85	normal
1271	(mul, mul, True, 0.75, 0.25, 0.25, 0.25, 7)	22010.56	normal
466	(add, mul, True, 0.5, 0.25, 1, 0.25, 7)	22291.86	normal
...	...	...	...
131	(add, add, True, 0.5, 0.5, 1, 0.5, 7)	29199.35	lr
368	(add, add, False, 0, 0.25, 0.25, 1, 7)	29482.36	lr
88	(add, add, True, 0.25, 0.25, 1, 0.25, 7)	29591.19	lr
254	(None, add, False, 0, 0.75, 0, 0.25, 7)	29602.09	lr
118	(add, add, True, 0.25, 0.75, 0.25, 0.25, 7)	29978.46	lr

	changepoint_prior_scale	seasonality_prior_scale	seasonality_mode	growth	changepoint_range	rmse
223	0.50	1.00	multiplicative	linear	0.95	49020.46
215	0.50	0.10	multiplicative	linear	0.95	49243.16
231	0.50	5.00	multiplicative	linear	0.95	49321.26
239	0.50	10.00	multiplicative	linear	0.95	49367.18
207	0.50	0.01	multiplicative	linear	0.95	50941.27
...	...	...	...	...	...	...
188	0.25	5.00	multiplicative	linear	0.10	341781.53
212	0.50	0.10	multiplicative	linear	0.10	342799.79
220	0.50	1.00	multiplicative	linear	0.10	342953.10
228	0.50	5.00	multiplicative	linear	0.10	342954.83
236	0.50	10.00	multiplicative	linear	0.10	342960.73

	0	1	2	3	4	rmse_crossval	rmse2_crossval	rmse_manual	rmse2_manual	crossval_avg	crossval_diff	manual_avg	manual_diff	diff-diff
24	0.10	0.01	multiplicative	linear	0.95	120940.54	121636.73	225410.62	203422.39	121288.64	696.19	214416.50	21988.23	22684.42
22	0.10	1.00	multiplicative	linear	0.95	109998.06	88989.14	215865.00	210831.44	99493.60	21008.92	213348.22	5033.56	26042.47
20	0.10	5.00	multiplicative	linear	0.95	109049.90	87664.34	209887.89	217701.60	98357.12	21385.56	213794.74	7813.71	29199.27
23	0.10	10.00	multiplicative	linear	0.95	110067.24	88198.10	215699.02	207822.46	99132.67	21869.13	211760.74	7876.56	29745.69
21	0.10	0.10	multiplicative	linear	0.95	109274.82	87542.99	219135.53	209457.12	98408.90	21731.83	214296.32	9678.41	31410.24
16	0.25	1.00	additive	linear	0.95	97216.10	68061.10	149004.15	230229.67	82638.60	29155.00	189616.91	81225.51	110380.51
15	0.25	5.00	additive	linear	0.95	96911.22	68296.25	146174.57	229921.48	82603.74	28614.97	188048.02	83746.91	112361.87
19	0.25	0.01	additive	linear	0.95	98697.63	69506.19	152084.91	237470.03	84101.91	29191.44	194777.47	85385.13	114576.57
18	0.25	10.00	additive	linear	0.95	98005.19	67511.87	147107.85	233005.54	82758.53	30493.32	190056.70	85897.69	116391.01
17	0.25	0.10	additive	linear	0.95	97405.96	68273.48	142957.77	231406.75	82839.72	29132.47	187182.26	88448.98	117581.46
31	0.50	5.00	multiplicative	linear	0.50	135545.01	107568.27	232376.26	124375.08	121556.64	27976.73	178375.67	108001.19	135977.92
29	0.50	1.00	multiplicative	linear	0.50	135474.37	107882.57	233641.84	124176.23	121678.47	27591.80	178909.04	109465.61	137057.41
30	0.50	10.00	multiplicative	linear	0.50	135482.70	107568.77	233421.47	123436.92	121525.74	27913.93	178429.20	109984.55	137898.48
28	0.50	0.10	multiplicative	linear	0.50	135445.16	107505.68	233326.77	122761.67	121475.42	27939.48	178044.22	110565.10	138504.58
35	0.25	5.00	multiplicative	linear	0.50	137383.01	112299.95	242127.99	128173.15	124841.48	25083.05	185150.57	113954.84	139037.90
32	0.25	0.10	multiplicative	linear	0.50	137338.01	112699.90	241753.47	126709.99	125018.96	24638.11	184231.73	115043.48	139681.60
39	0.50	5.00	additive	linear	0.50	138440.86	108597.23	223527.78	113021.18	123519.05	29843.64	168274.48	110506.60	140350.24
38	0.50	10.00	additive	linear	0.50	138424.11	108733.77	224114.55	113330.95	123578.94	29690.34	168722.75	110783.60	140473.94
34	0.25	1.00	multiplicative	linear	0.50	137377.58	112177.89	242611.65	127236.24	124777.73	25199.69	184923.95	115375.41	140575.10
36	0.25	10.00	multiplicative	linear	0.50	137388.87	112508.85	243485.25	127620.53	124948.86	24880.02	185552.89	115864.72	140744.74
37	0.50	1.00	additive	linear	0.50	138373.30	108691.74	224143.86	113074.68	123532.52	29681.57	168609.27	111069.19	140750.75
5	0.25	1.00	multiplicative	linear	0.95	78758.13	54408.77	120016.64	243846.13	66583.45	24349.36	181931.38	123829.49	148178.85
9	0.25	0.01	multiplicative	linear	0.95	80213.47	56155.48	120756.92	246711.93	68184.48	24057.99	183734.42	125955.01	150012.99
8	0.25	5.00	multiplicative	linear	0.95	79151.50	54642.49	118585.41	245465.09	66896.99	24509.02	182025.25	126879.67	151388.69
6	0.25	10.00	multiplicative	linear	0.95	78759.55	54522.53	115929.07	243643.38	66641.04	24237.01	179786.23	127714.30	151951.32
14	0.50	0.01	additive	linear	0.95	86569.84	58843.48	117301.87	241953.37	72706.66	27726.36	179627.62	124651.50	152377.86
7	0.25	0.10	multiplicative	linear	0.95	78924.09	54170.30	116147.38	244424.28	66547.19	24753.80	180285.83	128276.90	153030.70
11	0.50	10.00	additive	linear	0.95	85422.69	57592.37	115539.08	242059.15	71507.53	27830.31	178799.11	126520.08	154350.39
13	0.50	1.00	additive	linear	0.95	85463.79	57775.42	115484.50	242395.67	71619.61	27688.37	178940.08	126911.18	154599.54
12	0.50	0.10	additive	linear	0.95	85459.28	57443.62	113083.64	242646.80	71451.45	28015.66	177865.22	129563.16	157578.83
10	0.50	5.00	additive	linear	0.95	85405.80	57632.71	112995.73	243082.30	71519.25	27773.09	178039.01	130086.57	157859.66
4	0.50	0.01	multiplicative	linear	0.95	69628.07	50961.68	95477.54	252636.39	60294.87	18666.39	174056.96	157158.85	175825.23
2	0.50	0.10	multiplicative	linear	0.95	68628.26	49542.07	95822.84	252793.44	59085.16	19086.19	174308.14	156970.59	176056.78
3	0.50	10.00	multiplicative	linear	0.95	68942.85	49356.75	95689.37	252832.63	59149.80	19586.10	174261.00	157143.26	176729.36
0	0.50	5.00	multiplicative	linear	0.95	68496.52	49379.81	95417.64	253435.79	58938.17	19116.71	174426.72	158018.15	177134.86
1	0.50	1.00	multiplicative	linear	0.95	68536.12	49468.19	94651.58	253492.30	59002.15	19067.92	174071.94	158840.71	177908.64
33	0.00	0.10	multiplicative	linear	0.95	137342.49	195438.90	352065.57	224738.36	166390.70	58096.40	288401.97	127327.21	185423.61
27	0.00	10.00	multiplicative	linear	0.95	135015.32	194608.13	352951.07	224705.77	164811.73	59592.81	288828.42	128245.30	187838.11
25	0.00	5.00	multiplicative	linear	0.95	134317.58	195005.00	352942.53	224710.25	164661.29	60687.42	288826.39	128232.28	188919.70
26	0.00	1.00	multiplicative	linear	0.95	134429.62	196573.72	351844.35	224709.16	165501.67	62144.11	288276.75	127135.19	189279.30

CORONAVIRUS FUTURE PREDICTION¶

Sources of Data¶

0. Preprocessing¶

1. EDA¶

2. Machine Learning¶

3. Citations¶

Exploratory Data Analysis class=jump>[Jump to Beginning](#top)¶

0. INTRODUCTION [Jump to EDA Start](#eda)

Important Note:

1. WORLD [Jump to EDA Start](#eda)

2. CONTINENTS [Jump to EDA Start](#eda)

3. COUNTRIES IN CONTINENTS [Jump to EDA Start](#eda)

4. COUNTRIES IN CONTINENTS (BI-VARIATE ANALYSIS) [Jump to EDA Start](#eda)

4.1. Populace and Country Infrastructure¶

4.2. Government Measures for Pandemic¶

4.3. Country Indicators/Metrics¶

Life Expectancy¶

Population Density¶

GDP Per Capita¶

Literacy Rate¶

Democracy Index¶

Human Development Index¶

4.4. Mobility Trends¶

5. COUNTRIES [Jump to EDA Start](#eda)

6. SELECTED COUNTRIES [Jump to EDA Start](#eda)

6. BENFORD'S LAW [Jump to EDA Start](#eda)

Benford on Complete Dataframe</span>¶

Observations:¶

Benford on Individual Countries¶

Machine Learning [Jump to Beginning](#top)

0. INTRODUCTION [Jump to ML start](#ml)

Identifying Stationarity¶

Augmented Dickey-Fuller (ADF) test¶

Kwiatkowski–Phillips–Schmidt–Shin (KPSS) test¶

Identifying Seasonality¶

1. Seasonal Auto-Regressive Integrated Moving Average (SARIMA) [Jump to ML start](#ml)

Zero differencing¶

Single Order Difference (Non-seasonal)¶

Single-order Seasonal difference (Weekly)¶

Seasonal difference + Trend difference¶

Box-Cox Transformation¶

Detrend by Model Fitting (Linear Regression)¶

Grid Search¶

[JUMP](#skip_sarima_grid)¶

auto_arima¶

[JUMP](#skip_sarima_val1and2)¶

Test¶

Ensemble¶

[JUMP](#skip_sarima_ensemble)¶

Future¶

2. Triple Exponential Smoothing (Holt Winters) [Jump to ML start](#ml)

Grid Search¶

[JUMP](#skip_hw_grid)¶

Test¶

Ensemble¶

Future¶

3. PROPHET [Jump to ML start](#ml)

Basic out-of-the-box Prophet model with default parameters¶

Prophet model with manual parameters¶

Rolling Forecast¶

Grid Search¶

[JUMP](#skip_prophet_grid)¶

Top performing model from grid-search¶

Picking top 40 models¶

[JUMP](#skip_prophet_rmses2)¶

Test¶

Ensemble¶

Future¶

4. All Models (+ Ensemble) [Jump to ML start](#ml)

Combining the three models (Ensemble)¶

Combining SARIMA and Holt Winters (excluded Prophet)¶

Citations [Jump to Beginning](#top)

0. Preprocessing ¶

1. EDA ¶

2. Machine Learning ¶

3. Citations ¶